ABSTRACT. Uncertainty, exploration, and generalization in experience-based decisions

In many real-life decision problems, people need to experience the outcomes of their actions and learn which actions are most rewarding. This leads to an exploration-exploitation dilemma: do you choose the action which you currently think is most rewarding (exploitation), or a more uncertain action in order to learn more about it (exploration), potentially helping you obtain better outcomes in the future? Normatively, uncertainty should be a determinant of whether to explore or exploit. But whether this is the case for humans is still debated. We focus here on "contextual bandit" tasks, where there are noisy cues towards reward. In such tasks, clever exploration can help generalize prior experience to novel options, increasing the value of exploration. We propose a flexible Bayesian framework of function learning, called Gaussian process regression, to describe people's estimates of reward and the uncertainty associated to those estimates. To predict people's choices, we combine this with an uncertainty-guided decision rule. In three experiments (n=81, n=80, n=80), we find that exploration and generalization are indeed guided by uncertainty, consistent with our Bayesian account of learning and decision making.

ABSTRACT. When making a judgment it has been argued that people select among two kinds of judgment strategies: A knowledge abstraction strategy assumes that people learn to weigh over time which features reliably predict the judgment criterion, whereas a memory-based retrieval strategy proposes that people recall previous experiences from memory. To disentangle these strategies, past research has usually assumed that people consistently pursue the same strategy over time. The question of how strategies change as a function of learning has received less attention. In the current work, we formulate a learning model that develops a preference for one strategy over the other during learning by adjusting the relative importance of different features and past exemplars. Importantly, the learning model allows disentangling trial-by-trial strategy shifts from a global preference for integrating knowledge abstraction with memory retrieval. We test the models’ distinct predictions for judgment accuracy and familiarity-based choices in a judgment task requiring knowledge abstraction and memory retrieval. In line with the idea that people develop a global preference for one strategy, we find that people learn to accurately judge objects consistent with feature-based knowledge, whereas objects similar to previously encountered instances are more difficult to judge. In particular, new instances requiring retrieval are not more familiar to participants than new instances requiring feature-based knowledge, ruling out the possibility that participants shifted between strategies on a trial-by-trial basis. In sum, a learning model integrating knowledge abstraction and memory retrieval may provide a suitable tool for understanding learning processes in human judgment.

ABSTRACT. CALM – A process model of category generalization, abstraction and problem-structuring CALM is a category learning model, which integrates strong assumptions about the interaction of attention, generalization, rule abstraction and knowledge partitioning. Simulations show how CALM successfully predicts a variety of behavioural phenomena in differently structured categorization tasks, which were so far unexplained, or only accounted by qualitatively different models. CALM implements four hypotheses in an associative learning mechanism. First, associative learning is mainly dimensional (not configural), including similarity-based generalization from observed, and dissimilarity-based (inverse) generalization from unobserved instances. Second, paying attention to subjectively predictive (diagnostic) dimensions reduces learning about non-predictive dimensions. Third, strong beliefs from the past are sustained, and categorization errors do not lead to a direct adjustment of represented dimension outcome-relationships (self-affirmation). Fourth, such ‘obstinate beliefs’, however, allow to detect whether consequent categorization errors systematically occur in different environmental contexts, leading to behavioural adaptation by augmenting contextual modulation on existing expectations. An empirical test of the fourth hypothesis is presented. In two conditions participants learned to categorize objects with three binary attributes, with one of the attributes being predictive of category membership. After some learning, the category structure changed without announcement to an ‘Exclusive-Or’ structure (identical in both conditions). However, the learned structure - before - the change either could (condition A) or could not (condition B) be applied conditional on a context cue after the change. As predicted, learning the second category structure was significantly faster in condition A, compared to condition B.

ABSTRACT. The OpenModels project in category learning: Progress through distributed collaboration

Formal modelling in psychology is failing to live up to its potential due to a lack of effective collaboration. As a first step towards solving this problem, we have produced a set of freely-available tools for distributed collaboration. In this talk, I'll describe those tools, and the conceptual framework behind them. I'll also provide concrete examples of how these tools can be used within the study of category learning. The approach I propose enhances, rather than supplants, more traditional forms of publication. All the resources for this project are freely available from the catlearn website - http://catlearn.r-forge.r-project.org

ABSTRACT. Top-down and bottom-up guidance of visual attention during categorization decision making

When making categorization decisions, people look at spatial locations on a screen that have been associated with information about previous category members (exemplars), even when the exemplars are not visible anymore and have to be retrieved from memory. In this study, we systematically investigated the interaction of this so-called „looking at nothing “-phenomenon (LAN) with the accessibility of information stored in memory and with the presentation format during testing. In two experiments (N1 = 79, N2 = 54), participants repeatedly decided whether to invite job candidates for interviews in a multi-attribute categorization task. In one condition, participants learned the exemplars’ attributes by heart before learning to classify the exemplars, in the other conditions, the exemplars’ attributes were learned incidentally during the classification training. Additionally, we varied whether exemplars were presented visually or auditorily (Study 1) and how long the visual presentation lasted (Study 2). LAN during the categorisation of new test candidates occurred when exemplars were learned by heart and when they were only presented ten times on screen. LAN also occurred when presenting information auditorily and when removing visual information from the screen after 1.5 seconds. However, when visually presenting new test items without removing cue information, LAN did not occur. The results shed light on the interaction between visuospatial attention and attention to information in memory during categorization decision making as well as on the memory strength needed to elicit eye movements to emptied exemplar locations.

ABSTRACT. Modelling eye tracking dynamics with quantum theory

There have been several reports of how the structure of eye fixations leading to a simple (binary) decision can be predictive of the actual decision (e.g., the gaze cascade effect). However, there has been limited effort to directly model eye tracking curves. Part of the problem is that existing dynamical formalisms (such as multivariate decision field theory), which could be adapted for the modelling of eye tracking curves, provide limited scope for multiple reversals in weight for one vs. another option. We created binary decision task which is complex in the sense that the attributes/ advantages for each of the two options are not well-matched, with a view to foster multiple changes in attentional focus. We recorded eye fixations up to the decision point and confirmed the existence of multiple changes in attentional focus (which in itself is not surprising). We subsequently employed an open systems dynamical model from quantum theory, with a view to examine whether such a framework could describe eye-tracking curves. This was indeed the case and, moreover, there are indications that the quantum framework can reveal structure in the eye tracking dynamics predictive of the eventual decision of the participants. Finally, we consider the explanatory use of such a quantum framework for eye tracking dynamics.

ABSTRACT. Executive processes in the aesthetic appreciation of paintings

In this study we sought to determine the role of visual working memory in the appreciation of paintings by employing a dual task computer-based paradigm. Participants were asked to rate their liking for 48 paintings on a 5- point Likert scale, while simultaneously performing a visual working memory task. Half of the paintings were simple, and half were complex, as defined based on a pre-study. In each trial, participants first saw a 4 x 4 grid containing 1, 3, or 5 dots for 750 ms. These corresponded to the low, intermediate, and high working memory conditions. Thereafter they were asked to rate how much they liked an artwork presented for 2000 ms, after which they could use the mouse to submit their rating. After the rating had been completed, participants saw an empty 4 x 4 grid and asked to use the mouse to click the cells that contained the dots they saw before the artwork. Each click produced a dot in the matrix. In order to control for individual differences, participants subsequently performed a visual working memory span task, and completed questionnaires on openness to experience, and art interest, knowledge and activities. Results showed that liking ratings for artworks in the intermediate and high working memory load conditions were significantly higher than the ratings in the low working memory condition, both for simple and complex images.

ABSTRACT. Cross-cultural effects in aesthetics?

Cross-cultural studies in aesthetics are relatively rare, but it is generally assumed that cross-cultural differences should exist since cultures so clearly differ, although there is much equivocation, and little solid empirical data. The 1997 Handbook of Cross-Cultural Psychology showed the problems when it said, “The existence of both substantial differences and substantial similarities in the aesthetic responses of different cultures is largely self-evident”. Elsewhere in psychology there is currently a growing interest in cross-cultural cognitive differences, particularly following the influential work of Masuda and his colleagues, who have argued that 'East Asians' are more holistic and context-oriented in their perceptions than ‘Westerners' who are more analytic and object oriented, resulting in preferences for different types of photographs and art works. In this paper I will describe two studies, the first of which used the method of paired comparisons to assess differences in preference for a wide range of aesthetic stimuli, with 170 participants from the UK, Thailand and Hong Kong. The second study, with 146 participants, used a more focussed set of stimuli, as well as some from the first study to allow assessment of the reliability of preferences. Overall there was little evidence for consistent cross-cultural differences, which raises a number of theoretical questions.

ABSTRACT. The multidimensional nature of the preference for smooth curvature

The aesthetic appeal of curves has inspired experimental studies since the early 20th century. Up to nowadays, we report a consistent preference response for curvature (smooth vertices), as opposite to angularity (or “sharpness”). This has been found for a variety of visual stimuli: familiar and unfamiliar objects; abstract, irregular shapes; complex interior design environments and architectural façades. Recently, the origin of this phenomenon has been presented in relation to biological and culturally-based factors, hence suggesting its multidimensional facets. Using explicit tasks, we showed that type of response (rating scale vs. forced choice) did not modulate liking. The use of implicit tasks revealed an automatic association of curved shapes with positive (or “safe”) concepts and angular shapes with negative (or “dangerous”) concepts. However, angular shapes did not elicit any affective avoidance reaction, whereas curved shapes triggered approach. Beside the affective component, more recently, explicit preference for abstract curvature has been found in relation to art expertise and openness to experience. Taken together these results support the multidimensional character of preference for curvature, where the interrelation between bottom-up and top-down processes might be crucial.

ABSTRACT. Poetic speech melody

Research on the interface of music and poetic language has extensively investigated similarities and differences of musical meter, poetic meter, and the rhythm of ordinary speech, but not considered a potential counterpart of musical melody in the language of poetry. Moreover, linguistic research on melodic contours in natural language has remained limited to single phrases and specific speech acts, such as the rising prosodic contour of questions. Our study on poetic speech melody is the first to go beyond these limitations. We phonetically analyzed recitations of entire poems by multiple speakers to trace potentially converging contours of pitch and duration values across all stanzas that could be interpreted as inherent textual properties. Results show that stanzas of individual poems indeed feature––across a multitude of recitations––distinct recurrent pitch and duration contours, just like genuine songs and other pieces of music. Moreover, the higher the poem-specific autocorrelation score was for melodic recurrence across stanzas, the higher were also the ratings for subjectively perceived melodiousness as spontaneously reported by listeners. These ratings are also strongly predictive of beauty-attributions and overall liking.

ABSTRACT. The Aesthetic Homunculus: embodiment and expertise effects in aesthetics judgements

The term ‘embodied aesthetics’ has recently been coined to describe internal processes relating our own body and the observation of art. Our brain holds multiple representations of our body, here we explore how the somatosensory cortices (that holds a detailed sensory representation of our body) participates in visual aesthetic judgment of performing arts/dance. We performed an event-related potential (ERPs) study to test for the influence of visual aesthetic perception on somatosensory evoked potentials (SEPs). We recorded EEG in two groups of participants (expert dancers, control non-dancers) during the observation of pairs of body dance postures and whilst performing two tasks: (a) an aesthetic preference task, (b) a perceptual control task. The results showed differences during the aesthetic and perceptual tasks started at an early stage of processing (P45, N80, P100), and likely originated in primary and secondary somatosensory cortex. This early processing in somatosensory cortex was also modulated by the level of experience that the observers had with the observed dance posture. Overall, these results provide direct evidence for a differential involvement of somatosensory areas in the two task, suggesting different levels of embodiment during visual aesthetic judgement and a control visual judgment of a dance posture.

ABSTRACT. Do people like what the brain likes? The case of visual symmetry Within the long history of empirical aesthetics, a recurring theme is to what extent we can explain human preferences in terms of what is most salient, what attracts attention, and what excites a strong activation in sensory areas of the brain. A visual property often mentioned in this context is symmetry. There is clear evidence that extrastriate visual areas are activated by the presence of symmetry in an image (for a review: Bertamini et al. 2018). Some types of symmetries are more salient and generate a stronger response. Reflection symmetry in particular is both more salient and preferred. We found good overall correlation between preference and brain activity and a robust pattern across cultures (comparing participants in Britain and in Egypt, Makin et al., 2017; Bode et al., 2017). However, there are also two aspects that do not support a direct link. The first is the extent by which rotation symmetry is liked. The second is the effect of density of elements, which does not affect neural responses but can affect preference. Overall, the evidence is that in general we can say that people like what is salient and is processed efficiently in the brain, but other factors also influence preference.

ABSTRACT. Cognitive processes in risky choice under requirements Risky choice under requirements involves situations where people need to collect enough resources in a limited number of trials to meet a resource requirement. Each trial offers a low-risk and high-risk option with full description (risk as variance). Human decisions in such tasks partly follow the optimal dynamic programming solution to the problem (Houston & McNamara, 1988), which posits that the risky option becomes more preferable with harder requirements. This optimal solution, however, is computationally complex, considering every possible future state given every choice, outcome, and probability. We tested two cognitive models. The first model is a modified version of cumulative prospect theory; which has been proposed (McDermott, Fowler, & Smirnov, 2008) and criticized (Houston, Fawcett, Mallpress, & McNamara, 2014) in the recent theoretical literature. Our results, using predictive model tests, show that prospect theory predicts most participants’ behavior better than the optimal model. The second model is a decision tree with less compensatory comparisons than prospect theory and more attention to outcome magnitudes. The heuristic-like tree also predicts participants’ choices better than the optimal model. Comparing prospect theory to the decision tree reveals that prospect theory outperforms the tree in individual strategy classification, but not for the entire sample. Furthermore, a normative analysis revealed that prospect theory and the decision tree reach at least 80% of the optimal model’s expected value, showing that both models can perform near-optimal. This work shows that humans use highly-performant cognitive strategies in risky choices under requirements.

ABSTRACT. Probability Weighting: A Reflection of Option-Specific Attentional Biases During Evidence Accumulation?

An important theoretical concept for capturing phenomena in decisions under risk (e.g., certainty effect, fourfold pattern of risk attitude) is the notion of nonlinear probability weighting. The impact of the possible outcomes of a risky option on the decision can be described by assuming a systematic distortion of the outcomes’ probabilities (such that, for instance, outcomes with small probabilities are over- and large probabilities underweighted). It is currently unclear, however, whether and how patterns in probability weighting reflect distinct characteristics of information processing. We demonstrate that patterns of probability weighting can arise from attentional biases in a drift diffusion process, where preferences are constructed in an evidence accumulation process, biased by attention to the different options. We first simulated choices between a safe and a risky option with the attentional drift diffusion model, systematically varying the asymmetric attention to the risky vs. the safe option. The choices were then modeled with cumulative prospect theory. Higher attention toward the safe option during evidence accumulation was reflected in a lower elevation of the weighting function. In addition, asymmetric attention—irrespective of whether favoring the risky or the safe option—increased the curvature of the weighting functions. These findings suggest that the apparent over- and underweighting of probabilities might reflect asymmetries in option-specific attention, rather than distortions in the processing or representation of probability information. Moreover, our analysis establishes the mapping between two influential but previously unconnected computational models of decision making under risk.

ABSTRACT. It’s new, but is it good? How generalization and uncertainty guide the exploration of novel options

How do people decide whether to try out novel options as opposed to tried-and-tested ones? We argue that they infer a novel option's reward from contextual information learned from functional relations and take uncertainty into account when making a decision. We propose a Bayesian optimization model to describe their learning and decision making. This model relies on similarity-based learning of functional relationships between features and rewards, and a choice rule that balances exploration and exploitation by combining predicted rewards and the uncertainty of these predictions. Our model makes two main predictions. First, decision makers who learn functional relationships will generalize based on the learned reward function, choosing novel options only if their predicted reward is high. Second, they will take uncertainty about the function into account, and prefer novel options that can reduce this uncertainty. We test these predictions in two preregistered experiments (N=320 and N=423) in which we examine participants' preferences for novel options using a feature-based multi-armed bandit task in which rewards are a noisy function of observable features. Participants' expectations, confidence and choices reveal strong evidence for functional exploration and moderate evidence for uncertainty-guided exploration. In contrast to previous explanations, these results indicate that reactions toward novelty are guided largely by fine grained contextual knowledge, with uncertainty-guided exploration playing a smaller role.

ABSTRACT. Remembering retrieval: Evidence from repeated recall

Observations about the beneficial effects of retrieval for subsequent memory date as far back as the writings of Aristotle, and there is a large body of modern research investigating the impact of retrieval practice on successful recall. However, models of this phenomenon typically focus on what factors can explain retrieval probability, rather than whether and how memory representations themselves change. I conducted a series of experiments (total N=231) aimed at addressing these latter issues by asking participants to immediately and repeatedly recall colour information on a continuous response scale (i.e., a colour-wheel). Results showed a gradual, decelerating deterioration of recall performance, coupled with a reduction in recall variability, as participants made more recall attempts. This pattern was also present when the majority of retrievals did not require an overt response, suggesting it results from memory rather than response processes. Models in which recall attempts produce novel memory traces alongside the original memory provided a better account of the data than models in which traces overwrite their predecessors, or models in which old and new information is merged into a single memory trace. This is consistent with a recent theory of retrieval practice (Rickard & Pan, 2017), and suggests that retrieval from immediate and long-term memory has similar implications for memory representations.

ABSTRACT. Age differences in hindsight bias: A meta-analysis with a Bayesian multinomial processing tree approach

After people learned a fact or the outcome of an event, they often overestimate their ability to have known the correct answer beforehand. This hindsight bias has two sources: an impairment in direct recall of the original (i.e., uninformed) judgment after presentation of the correct answer (recollection bias) and a reconstruction of the original judgment that is biased towards the correct answer (reconstruction bias). Research on how cognitive aging affects the sources of hindsight bias produced mixed results. To synthesize available findings, we conducted a meta-analysis of nine studies (N = 366 young, N = 368 older adults). We isolated the probabilities of recollection, recollection bias, and reconstruction bias with a Bayesian, three-level hierarchical implementation of the multinomial processing tree model of hindsight bias (Erdfelder & Buchner, 1998). Additionally, we quantified the magnitude of bias in the reconstructed judgment. Overall, older adults were less likely to recollect their original judgment than young adults, and thus had to reconstruct it more frequently. Importantly, outcome knowledge impaired recollection of the original judgment (i.e., recollection bias) to a similar extent in both age groups, but outcome knowledge was more likely to distort reconstruction of the original judgment (i.e., reconstruction bias) in older adults. In addition, the magnitude of bias in the reconstructed judgments was slightly larger in older than in young adults. Our results provide the basis for a targeted investigation of the mechanisms driving these age differences.

ABSTRACT. A Bayesian and Frequentist Multiverse Pipeline for MPT models – Applications to Recognition Memory

Even with a clear hypothesis or cognitive model in mind, most statistical analyses contain several more or less arbitrary choices. In the case of a model-based analysis, these choices can concern the statistical framework, the aggregation-level, and which parameter restrictions to introduce. Usually one path through this ‘garden of forking paths’ (Gelman & Loken, 2013) is chosen and reported. However, it is unclear how much each choice affects the reported results. The multiverse approach (Steegen, Tuerlinckx, Gelman, & Vanpaemel, 2016) offers a principled alternative in which results for all possible combinations of reasonable modeling choices are reported. We developed a software package for R that performs a model-based multiverse analysis for multinomial processing tree (MPT) models, MPTmultiverse. Our package estimates MPT models in a frequentist and Bayesian manner. In the frequentist case, it uses no pooling (with and without bootstrap) and complete pooling. In the Bayesian case, it uses no pooling, complete pooling, and three different variants of partial pooling. We applied our approach to a large confidence-rating recognition memory data corpus consisting of 12 studies with over 450 participants using a relatively unrestricted variant of the 2-high threshold model for confidence ratings (Bröder, Kellen, Schütz, & Rohrmeier, 2013). Our results show that even for some core parameters, the different analysis approaches reveal considerable variability in the parameter estimates across estimation methods. Our results suggest that researchers should adopt a multiverse approach when using cognitive models.

ABSTRACT. During natural behavior – in comparison to laboratory investigations – we rarely use cognitive subsystems (e.g. WM) to capacity or follow a strict top-down encoding protocol (e.g. explicit memorization). In fact, we often act short-term task-oriented, which results in a behaviorally optimal representation of our environment. In a series of computer-based, virtual-reality, and real-world experiments we investigated the role of sampling information during natural tasks (via eye movements and incidental exposure durations) for the generation of visual long-term memories (VLTMs). Even after incidental encounters with thousands of isolated objects, VLTM capacity for these objects is reliable, however, the detail of these representations is quite sparse and does not increase if the incidental encounters are longer(Study1). When searching for objects in an actual real-world environment – where locomotion is necessary – fixation durations on objects predict subsequent location memory, as measured with surprise memory tests(Study2). Incidental representations generated during search are more reliable than memories established after explicit memorization in a realistic virtual environment(Study3). Further, in a real-world object-sorting task, eye movements used for minimizing WM load and instead gathering task-relevant information just before it is required, significantly predict long-term location memory of objects (Study4). Finally, in a virtual-reality paradigm (Study5), we show that spatial priors can be activated within the first fixation into a new environment in a one-shot manner. Together, this rich set of studies shows that incidental information acquisition during natural behavior establishes reliable VLTM representations, which can be used to guide ongoing behavior in a proactive fashion.

ABSTRACT. In two experiments we investigated the effect of pitch scale stability on the implicit learning of an artificial musical structure. In a first experiment, participants (N=20) were exposed to melodies generated by an artificial grammar under incidental learning conditions. For one group of participants, the underlying pitch scales were always the same for each melodic realization (stable). In another group, each melody was realized on randomly detuned versions of an abstract target scale (unstable). In a subsequent test phase grammatical melodies had to be distinguished from ungrammatical melodies. The melodies could be realized on a detuned or on the stable/target scale. Both groups of participants performed equally well on this task for test melodies stemming from the stable/target scale. However, discrimination performance of subjects who only learned with stable melodies was significantly impaired for detuned melodies. On the other hand, participants who were exposed to detuned melodies from the first moment performed equally good on detuned and stable melodies (drawn from their target scale). In a second experiment (N=20), we used randomly generated pitch scales instead of detuned versions of an abstract target scale. We obtained similar results for both groups albeit the overall performance for random scale melodies were much more impaired compared to the slightly detuned melodies from the first experiment. Since melodic interval information is no longer available here, subjects apparently use a coarser contour code to solve this task.

ABSTRACT. Quizzing has been shown to be an effective means to foster learning. In the present study, we investigated if providing learners with quiz questions adapted to the cognitive load that they perceive when answering quiz questions (adaptive quizzing) would further increase the beneficial effects of quizzing found in the literature. Our study is based on research on cognitive load that shows that taking into account learners’ cognitive resources when constructing or choosing learning material (e.g. questions on a quiz) increases learning outcomes. We conducted an experiment following a one factorial between subjects design in which we compared the learning outcomes of N = 160 university students. Using a digital learning environment, participants first watched an e-lecture and then learned its content according to one of three randomly assigned conditions: (a) adaptive quizzing, (b) non-adaptive quizzing and (c) note-taking. Participants in the quizzing conditions answered open-ended questions on four levels of difficulty. In the adaptive quizzing condition, the difficulty of the questions depended on the cognitive load indicated by the participants; in the non-adaptive condition, the difficulty of the questions gradually increased. Participants in the control condition note-taking were asked to take notes on the e-lecture. Learning outcomes were measured one week later via a posttest. We found that adaptive quizzing yielded higher learning outcomes than non-adaptive quizzing. Both quizzing groups outperformed the note-taking group. We conclude that, in order to increase the yield of quizzing, the cognitive load imposed on learners by the quiz questions should be taken into consideration.

ABSTRACT. Introduction In multitasking predictability of a task is likely desirable because it enables automatic task control which is less dependent on cognitive resources. In two experiments, we tested the effect of predictability by letting participants practice a predictable tracking task and a go/no-go auditory reaction time task (as single-task or dual-task). Methods In the first experiment 33 participants practiced a tracking task and auditory reaction time task separately, each containing predictable sequences. For the second experiment 39 participants trained the tasks as a dual-task. In both experiments participants were divided into an implicitly and explicitly instructed group. Results Predictability improved tracking (Predictable: M = 1.46 cm, SD = 0.32 cm, Random: M = 1.55 cm, SD = 0.34 cm) and reaction times (Predictable: M = 426 ms, SD = 79 ms, Random: M = 567 ms, SD = 55 ms) during dual-tasking. However, predictability in one task did not improve performance on the other task. Results of the dual-task training experiment mirrored those of the single-task training experiment but dual-task training led to dual-task performance surpassing single task performance. In neither experiment group differences were found. Discussion Predictability has a positive but robustly unilateral effect on dual-task performance. A possible explanation is the process of task-shielding, whereby tasks are kept apart in order to reduce interference, but also has the negative effect of inflexible task processing where resources can’t be distributed flexibly when task demands might call for it.

ABSTRACT. According to ideomotor theory, when people observe an action being followed by a certain effect, they acquire bi-directional action-effect associations. At a later point, when they want to achieve the effect, its anticipation activates the corresponding action and facilitates its execution. The expression of action-effect associations is assumed to take place incidentally and automatically. However, verbal instruction and particular features of a task are known to influence action control mode and action selection. Here, we assess the impact of verbal instruction and the task relevance of action effects on the expression of action-effect associations in three experiments. Our results show that an effect-based instruction prompts participants to express action-effect associations more strongly than an action-based instruction. In addition, the acquired associations are only expressed when the effects are relevant for the task and when they are presented in the test phase. These findings show that in contrast to the assumptions of ideomotor theory, action-effect knowledge is only used when it is considered meaningful for the intended goal. Verbal instruction and task features affect the intentional state and may therefore bias the results of ideomotor studies.

ABSTRACT. We investigate inner-individual dilemmas (IIDs) in the context of personnel evaluation. In IIDs, reminiscent of social dilemmas, the optimization of local goals has effects on other goals (externalities) and thereby may lead to suboptimal results for the global outcome for an individual. We have previously begun to investigate Two-Level Personnel Evaluation Tasks (T-PETs), where participants were put in the role of (neutral) human-resource managers. Normatively, they should be interested in the overall outcome of work groups. They had to evaluate employees in settings where individual and group contributions could be dissociated. We first sketch previous results that suggested a “Tragedy of Personnel Evaluation” (von Sydow, Braus, & Hahn, 2018). When evaluating an ‘altruistically’ (or strategically group-serving employee) who had the smallest individual contribution, but by far the greatest positive effect on the group’s overall earnings, this employee was often rated the most negatively. In selection-tasks this employee was often excluded from the group. When we removed conflicting information, however, most participants were able to detect quickly the largest overall impact of the 'altruist'. Here we report two further T-PET experiments (N = 121; 172) that examine whether emphasizing the group by context or by variation of the level of data shown (successions of rounds with local, global, or both kinds of information) can avert ‘tragic’ outcomes. The results show some improvements and new perspectives, but also that the ‘Tragedy of Personnel Evaluation’ is quite robust.

ABSTRACT. A meta-stereotype is a judgment about how group members think that an out-group stereotypes their in-group. In order to capture meta-stereotypes, we use the Stereotype Content Model (SCM). The distinctive contribution the SCM makes is that stereotypes often hold mixed ascriptions of those two dimensions and assumes an incongruent structure in the warmth-competence space (i.e. being high on one dimension and low on the other) that lead to more extreme judgements. This is closely associated with previous research that indicates that judgements are made in a comparative context. In two studies, we investigate meta-stereotypes in a European setting. In the first study, we investigate how meta-stereotypes of two European nations depend on a comparative national context. Germans and Italians indicated their meta-stereotypes towards the target out-group nations Spain and Great Britain. These contexts were chosen because they exhibit an incongruent stereotype structures based on the Stereotype Content Model: either high in warmth and low in competence or vice versa. The results show that people overstate meta-stereotypes when the meta-stereotypical target is in an opposite position on the warmth-competence-dimension. In a congruent setting, people exhibit less meta-stereotyping. The second study is aimed to replicate the first findings with a different set of German and Spanish participants and expands this research into the field of working professions. In addition, possible moderators like the Social Dominance Orientation and Subjective Social Status of the participant are investigated. All results are discussed in terms of the role of comparative context in meta-stereotyping.

ABSTRACT. Nowadays, information seekers have the opportunity to use different virtual learning environments (e.g., online courses with video lectures and online information forums) to acquire knowledge and develop new skills. However, most of the Internet is not governed by editors or other gate-keeping institutions and therefore the question arises whom to trust and which information is credible. How do information seekers make such decisions? Using three 2x2 between-subject online experiments (N = 539), we test the hypotheses that the language style (experiment 1: neutral vs. aggressive; experiment 2: neutral vs. enthusiastic; experiment 3: neutral vs. praising) and the professional affiliation (experiments 1-3: scientist vs. lobbyist) of an information source influence the evaluation of (a) his trustworthiness and (b) the credibility of the provided information in virtual learning environments. Overall, analyses of variance show that the professional affiliation influenced the trustworthiness of the information source, but it did not influence the credibility of the provided information. More specifically, scientists were perceived as more trustworthy than lobbyists, but their information was not perceived as more credible. In contrast, all language style manipulations influenced the trustworthiness of the information source as well as the credibility of the provided information: When the information giver used an aggressive, enthusiastic, or praising language style, he was perceived as more manipulative, less benevolent, less honest and the provided information was perceived as less credible. Results are discussed regarding the communication of trust and related research.

ABSTRACT. Effect-based accounts of human action control have recently highlighted the possibility of representing one’s own actions in terms of its anticipated changes in the behavior of social interaction partners. In contrast to action effects that pertain to the agent’s body or the agent’s physical environment, social action effects have been proposed to come with peculiarities inherent to their social nature. Here we revisit the currently most prominent demonstration of such a peculiarity: the role of eye-contact for action-effect learning in social contexts (Sato & Itakura, 2013). In contrast to the previous demonstration of action-effect learning, a conceptual (N = 48) and a direct replication (N = 24) both yielded evidence for the absence of action-effect learning in the proposed design irrespective of eye-contact. These results suggest a limited generalizability of the original findings, either due to limitations that are inherent in the proposed study design or due to cultural differences.

ABSTRACT. How do individuals integrate discordant advice with prior estimates? This study tests a two-step model of individual belief updating. The first step determines whether individuals take advice into account or reject it completely. In the second step, individuals that decided to take advice integrate it. We hypothesized that the first step is influenced by whether the initial estimate is an outlier to the advice, while the second step is influenced by the distance between the mean of advice and the initial estimate. This model is tested in a context in which advice is provided by advisors that are a member of either the majority, the participant’s ingroup, or an outgroup perceived as competent. This allows to investigate the influence of majority conformity, ingroup conformity and group-level competence on both steps of belief updating. After collecting answers to 28 socio-demographic questions about the EU in a series of pre-studies (N = 120), these were presented as advice in four within-participant conditions to 40 British participants. The results of linear mixed-effects modelling support the two-step model of belief updating as superior to models of advice integration proposed by the literature. In addition, factoring in the influence of different social groups improved model performance in both steps through accounting for interindividual differences in the weighting of advice from different groups. We conclude that the two-step model of belief updating not only allows better predictions of updating, but also the opportunity to test the influence of advisor-, advice-taker- and advice-level influences on both steps.

ABSTRACT. In times of digitization, many scientific talks are recorded. Previous research has shown that people use trivial factors (e.g., sound quality) to evaluate the quality of scientific research. The good sound good research effect suggests that good audio quality leads to better perceived quality of the research and the researcher in comparison to bad audio quality. As the replication crisis has shown, such findings need to be replicated. We tried to replicate the effect (originally found in an US-American sample) with a student sample from Germany and a slightly different method. Our hypotheses were in line with the good sound good research effect. In experiment 1, participants (N = 157) listened to a science podcast with either good or bad audio quality. In experiment 2, participants (N = 161) watched a scientific conference talk with either good or bad audio quality. Afterwards, participants evaluated the respective researchers and their research. To test our hypotheses, we conducted independent t-tests. Results of experiment 1 show that participants give the podcast a better overall rating if it was presented with good audio quality. However, audio quality does not influence the perceived quality of the research and the competence of the researcher. Results of experiment 2 show that participants like the researcher more and rate his research as more important if it was presented with good audio quality. However, audio quality influences neither the competence of the researcher nor overall rating of the talk. Reasons for replicating the original findings just partly are discussed.

ABSTRACT. The role of processing fluency in judgments is undebated. Not only is fluency affected by sources as diverse as stimulus repetition or visual clarity, it also affects outcomes as diverse as aesthetic pleasure or subjective validity of a statement. While several studies indicate that fluency sources and outcomes are interchangeable, we propose that fluency effects are subject to a match between the source of fluency and the judgment dimension. Specifically, we propose that conceptual fluency is more informative for content-related judgments, but perceptual fluency is more informative for judgments related to perception. Three experiments (total N = 582) tested this hypothesis by implementing a paradigm that allows the orthogonal manipulation of conceptual vs. perceptual fluency, and the judgment of content vs. perceptual stimulus dimensions, respectively the truth value and the aesthetic appeal of statements. Conceptual fluency was manipulated by repeating the content of the statements. Perceptual fluency was manipulated through visual figure-ground contrast in which statements were presented. We predicted that judgments of truth would be more influenced by repetition than by visual contrast, and that judgments of aesthetic appeal would be more influenced by visual contrast than by repetition. Results of the experiments support these predictions, showing the superiority of content repetition on judgments of truth, but the superiority of visual contrast on aesthetic judgments.

ABSTRACT. The order of target and standard in a paired comparison not only determines the direction of the comparison (upward or downward) but also the efficiency of the comparisonprocess. We tested this in a series of 8 experiments (N = 856), where we let participants repetitively perform simple magnitude comparisons of two objects (e.g. one digit numbers or geometric shapes) in various economic and social contexts. Results showed that performance (speed and accuracy) of participants is facilitated when they encounter the standard stimulus before the to be judged target stimulus. A Standard-Target-Sequence Effect (STSE) on comparison processes so far was rather relevant in signal detection and stimuli discrimination tasks in psychophysics (so called Type B Effect, e.g. Dijas & Ulrich, 2014). Social and cognitive psychologists’ research on judgements of similarity and contrast have provided inconsistent results for the influence of the sequence of standard and target on the comparison process (e.g. Tversky, 1978; Angostinelli, Sherman, Fazio & Hearst, 1986). Our findings are a strong argument for an underlying mechanism in working memory that contradicts most lays' and experts' intuitive assumption: To perform a comparisons most efficiently one has to consider the standard before the target.

ABSTRACT. Research has shown that affective responses to a stimulus can influence risk judgements. The complex stimuli of multicue probabilistic judgements have numerous distinct features, each capable of evoking affect irrespective of relevance for the risk judgement. We first examine whether the overall affective response to a complex stimulus, measured as attractiveness, influences its risk judgement. We then investigate whether one feature of the stimulus, having different subjective affective values, influences the risk judgement despite its indeterminate relevance for that judgement. Across 3 studies (N=141), participants judged the safety of 4 backcountry skiing scenarios in terms of avalanche danger. The studies used a mixed design: the within-subject experimental factor was the presence vs. absence of ski tracks, the quasi-experimental factor was the degree of attractiveness of the scenario, and the main response variable was the safety judgement with respect to avalanche risk. Results indicate a positive relationship between scenario attractiveness and safety judgements: increasingly positive affective responses correspond to judgements of greater safety. This relationship holds for novice and expert backcountry skiers. The results further indicate a significant interaction between the preference for no ski tracks and the judgement of safety: the greater the individual preference for the absence of ski tracks, the safer scenarios with no ski tracks were judged to be. A sufficiently strong affective response to a single feature of a stimulus, irrespective of its relevance for the risk judgement, can significantly influence the risk judgement.

ABSTRACT. The ability to perceive structure, detect rules, and extract associations is a hallmark feature of human cognition. Yet people search for patterns even in random sequences—a tendency that has been argued to give rise to suboptimal decisions in repeated choice under uncertainty. In this study, we examined the decisions of collaborating three-person groups (n = 81) and independent individuals (n = 81) in a two-part repeated choice task. In the probabilistic part of the task, outcomes were serially independent and the optimal strategy was to exclusively select the most likely outcome (i.e., probability maximize). In the patterned part, a fixed outcome sequence was repeated and choice optimality was gauged from participants’ pattern accuracy (i.e., proportion of correct choices). We found that groups performed as well as the best individuals in the probabilistic part of the task. By contrast, groups’ pattern accuracy was not credibly higher than that of the average individual. Moreover, groups who did not identify an existing pattern, systematically underperformed by attempting to maximize probability. Qualitative coding of group discussion data revealed that failures to identify existing patterns were largely driven by a tendency to accept probability maximizing as a “good enough” strategy rather than to expend the effort of searching for a predictable outcome sequence. These results suggest that there are two routes to maximizing probability in repeated choice under uncertainty: recognizing that a probabilistic process cannot be outdone (maximizing as optimizing) or contenting oneself with an imperfect but easily implementable prediction strategy (maximizing as satisficing).

ABSTRACT. The large majority of research on multiple-cue judgment has focused on the decision strategies of individual decision makers and has relied on outcome measures to investigate inference mechanisms. By contrast, dyadic multiple-cue judgment, during which small groups of individuals are able to communicate freely, has received only little attention. In this study, we developed a qualitative coding scheme for classifying dyads’ discussion statements in a multiple-cue judgment task as indicative of either cue-abstraction or exemplar-based inference mechanisms. Based on the qualitative data from 40 dyads, we examined the effects of two different learning conditions (n = 20) on subsequent judgment accuracy and strategy use: learning by comparison, in which dyads learned which of two objects had a higher criterion value, and direct criterion learning, in which dyads learned each object’s criterion value directly. We found evidence for a social abstraction effect — during group discussions, dyads expressed more statements indicative of a cue-abstraction than an exemplar-based strategy. This effect was more pronounced after training in the learning by comparison condition. Furthermore, evaluation of judgment accuracy suggested better cross-item generalization and extrapolation for dyads classified as cue-abstraction users. Finally, comparing the strategy classification results obtained from qualitative coding with those of a computational modeling analysis, revealed a high consistency between the two approaches. These findings suggest that dyads’ verbalizations of decision strategies are consistent with the underlying cognitive processes and that qualitative coding of verbal protocols can provide valuable insights into the mechanisms underlying multiple-cue inference.

ABSTRACT. Recent studies claimed that time pressure can change risk preferences in decisions between risky gambles. Yet, often inference is based on group choice proportions. When choice proportions deviate from 50% in control, a significant change towards 50% could either be modeled as a change in preferences or an increase in noise. Without a stochastic choice model, one cannot distinguish between both hypotheses. In two studies (n=40 and n=60), we examined binary risky choices with and without time pressure in a within-subjects design. In each condition, people chose 120 times between two gambles that varied in expected value and variance. In addition, we varied whether the safer or riskier gamble was more complex. In the first study complexity was manipulated as gambles consisting of either one sure thing or a 50-50 gamble compared to two-outcome gambles with unequal prospects. In the second study, complexity was varied by presenting gambles with either 2 or 4 outcomes. Using Bayesian hierarchical estimation of utility functions with a probit link function, we showed that time pressure decreased consistency (leading to more noise), but did not systematically affect risk preference. Furthermore, we found evidence for complexity aversion in that people were less likely to choose the more complex gamble. This effect maintained controlling for expected value, variance, and consistency. We conclude that stochastic utility modeling is necessary to understand the effect of time pressure on risky choice and that gamble complexity has to be held constant to measure risk preferences.

ABSTRACT. Children’s sibling relationships and friendships form important developmental contexts, especially on the issue of moral development (Dunn 2014, Schmid/Keller 1998, Youniss 1994). Moral obligation and interpersonal responsibility matter crucially within relationships (Keller 1996). I supposed that children feel obligated to moral norms depending on the particular relationship and expected children to be stronger committed to their siblings than they are to their friends. To investigate these hypotheses, children in the 5 to 6 age group (N = 83) were asked questions concerning their siblings and friends in a semi-standardized interview. Second, they were presented 5 scenarios referring to positive duties (sharing, helping) and to negative duties (stealing, promise). The analysis (Qualitative Content Analysis and Chi-square tests) of the interviews revealed that sanctions, justice, welfare or norms were not striking for their decision. Children referred to their relationships and the moral commitment arising from these. Choosing their sibling as well as choosing their friend correlates significantly with their relationship. In addition, the results reveal that children are stronger committed to their siblings than they are to their friends. Especially the changes of perspectives within the scenario sharing a sweet clarifies this result. This pattern appears within positive duties (sharing, helping) and within the positively structured duty (promise). Stealing is only an issue within friendships. Taking something away is not classified as stealing within sibling relationships. As the results support the link between relationships and moral development, future research on moral development should take relationships more strongly into account.

ABSTRACT. According to the lateral attitude change (LAC) model (Glaser et al., 2015), a displacement effect occurs when people explicitly resist a persuasion attempt toward a focal object, but do change their attitude toward a lateral object, which was not the target of persuasion. Similar to the sleeper effect (Hovland, 1949), however, the reasons for resisting focal persuasion may become inaccessible, and thus a delayed explicit attitude change toward the focal attitude object may emerge. In an ongoing 3x2 experiment, participants read either one of two pretested persuasive messages (favoring or opposing inclusive schooling) or a non-persuasive control text. Immediately afterwards, they receive a disclaimer (or validation message) declaring message content invalid (valid). Participants' explicit and implicit attitudes toward the focal attitude object inclusive schooling and one lateral attitude object are assessed both immediately (T1) and after a one-week delay (T2). Results are expected to be in line with the LAC model: Participants who read no disclaimer will show a pattern of generalization (explicit attitude change toward focal and lateral objects) at both T1 and T2. Participants who do read the disclaimer will show a pattern of displacement (explicit attitude change only toward the lateral object) at T1, and of generalization at T2. Implicit attitudes will always show a generalization pattern, as they are unaffected by conscious rejection of the message. Results and implications for theory will be discussed.

ABSTRACT. Information search is greatly curtailed when competition threatens to pre-empt one’s choice. Inspired by Phillips, Hertwig, Kareev, & Avrahami (2014), we introduced a computerized rival into our speed-accuracy trade-off paradigm to encourage participants (N = 103) to make their binary sampling-based decisions more optimally than in previous studies. The trade-off arises from a limited total session time, in which the number of decisions possible depends on the time spent sampling. Since correct and incorrect decisions were rewarded and punished, there was a trade-off between two goals. On one hand, one ought to maximise the number of decisions one encounters to maximise the potential total payoff (speed component), on the other hand, one ought to maximise sample size and thereby accuracy and payoff per decision (accuracy component). While the optimal strategy demands a strong focus on the speed component, participants considerably overemphasise accuracy. Introducing a rival, who followed a close-to-optimal strategy and could pre-empt the participants’ decision proved successful in fostering a faster strategy and showed transfer effects in a subsequent standard session. However, our additional conditions, in which participants merely watched a computerised teammate execute a close-to-optimal strategy (either alone or playing against the rival) induced similar transfer effects. Moreover, all conditions regressed towards a slower strategy than what they had observed or applied. Learning from experience seemed no better than vicarious learning, and neither seemed to provide a suitable learning environment for participants to gain enough insight into the optimal strategy to uphold or even outperform the demonstrated strategy.

ABSTRACT. People often forget acquired knowledge over time such as the names of previous prime ministers. Judgments hinge on the knowledge people possess. For instance, the decision to vote for a party may rest upon a similarity-based retrieval of the previous political successes a party reached or a more rule-based assessment to what degree the party follows liberal positions. Past research has suggested judgment accuracy declines for the rule- and similarity-based judgments over time, but practice helps to restrict this decay for rule-based judgments. One potential reason why people forget is they learn new information that interferes with the previously acquired knowledge. The present study aims to further shed light on how forgetting processes limit the human ability to make accurate judgments by studying how learning of new information interferes with the judgment policies people have established earlier. Specifically, we expect that if participants have established a rule-based judgment policy, changing this rule will be more detrimental to their judgment policy than storing new knowledge in the form of single exemplars and vice versa. To test these predictions, participants learned to solve a judgment task in which they had to identify two exceptions among one linear, additive rule. In a second training phase, participants than either learned another rule, new exceptions, or both. A control group did unrelated task. In the last phase, participants were asked to remember how they judged the stimuli in the first phase. Taken together the study aims to better understand the consequences of forgetting in judgment.

ABSTRACT. The present paper investigates the influence of situational factors on subjective experiences of justice in school. The perceived justice of interactions is especially important in inclusive classrooms as it fosters the students’ sense of belonging and community. Based on a psychological perspective of justice, we developed vignettes focusing student-teacher-interactions. In an experiment we varied the special educational need of the student in 6 vignettes. A total of 275 teacher students of the University of Potsdam evaluated the perceived justice of the situations. Linear mixed effect model analyses showed a significant influence of the student’s special educational need on justice judgements, specifically for behavioral problems and marginally for learning difficulties. The diagnostic status did not influence the justice judgements. These findings suggest a shift from the principle of equality to the principle of need in the formation of justice judgements.

ABSTRACT. A current trend in human-robot interaction is to create social robots which resemble human beings. This can be done by anthropomorphic design or by framing a robot humanlike. The current study examined how framing a robot affected people’s perception of it as well as their willingness to bear financial costs for the robot’s sake. It was hypothesized that anthropomorphic framing would lead to a more humanlike perception of the robot and to greater willingness to give up money for it. Half of the participants (n = 20) received an anthropomorphizing story about NAO, the robot, and the other half (n = 20) read a mechanical description of it. All subjects were supported by NAO in solving the Tower of Hanoi, a mathematical game. Afterwards, participants filled in a questionnaire assessing perceived humanness, eeriness, and attractiveness of the robot and received 5€ for participation. When they were about to leave the laboratory, they were told a fake story about why NAO needed a repair. They were then asked if they wanted to donate an arbitrary amount of the money they just received for this specific purpose. Framing did not yield a significant difference in people’s perception of NAO. However, contrary to our hypothesis, people donated significantly less when the robot was framed humanlike. This suggests that anthropomorphic framing may induce people to dismiss the robot’s purpose as a tool by creating a context in which it is perceived more like an individual subject rather than a machine (object).

ABSTRACT. Recent theories on decision-making in adolescence suggest a peak of risk-taking in mid-adolescence, while actual findings imply that this might only be true for highly rewarding and social task contexts. The aim of this study was to examine whether differences in adolescent risk-taking might rather be explained by individual differences in impulsivity and empathy than age, raising doubts that the biological factor age is the most critical factor in understanding risk-taking in adolescence. To investigate this, we applied a STOPLIGHT task, a well-known task to measure risky decisions in a simulated driving game. Overall, 50 early- (9-11 years), 63 mid- (12-14 years), and 56 late-adolescents (15-17 years) were faced with the decision to either stop (safety-decision) or run over (go-decision) yellow traffic lights. With go-decisions, they saved time else spent at the traffic light but risked to cause an accident and to lose more time. Adolescents were told to progress as fast as possible to arrive at a friend’s party. We measured impulsivity and empathy with self-report questionnaires and separated each age group in high or low impulsive and empathic subjects. Results indicated (a) more go-decisions in high impulsive and high empathic participants irrespective of age, (b) lower go-decisions in low than high impulsive and empathic individuals only in early adolescents, and (c) low impulsive and empathic mid- and late-adolescents did not significantly differ in go-decisions from high impulsive or empathic peers. These findings suggest that individual differences in personality need to be considered in developmental theories on adolescent decision-making.

ABSTRACT. The religious background of refugees and their behavior towards women is a controversial issue in German politics, society, and media ever since the incidents of sexual assaults during New Year’s Eve in Cologne 2015/2016. As part of an online study, we examined the effect of the perpetrator’s ethnicity, negative attitudes against Muslims and the Islam as well as rape myth acceptance on participants’ perpetrator and victim related judgments of a rape case. In a between-subjects-design, 987 subjects read an ambiguous rape vignette in which the perpetrator’s ethnicity was manipulated by his name. As expected, rape myth acceptance predicted victim blaming and perpetrator exoneration, independent of the experimental condition. Contrary to our expectations, negative attitudes against Muslims and the Islam predicted more lenient judgements for the migrant perpetrator than for his German counterpart. We discuss possible boundary conditions of and processes underlying this leniency effect.

ABSTRACT. There is an ongoing debate whether adolescents are poor or reasonable decision-maker in experimental risk settings. As findings remain inconsistent, they are mostly limited to decisions when maximizing gains. The aim of this study was to investigate how the incentive valence (gain, loss) and expected values (EV) of safe and risky options influence adolescents` risk-taking. Based on findings in adults, we expected that more risks should be taken in loss than gain situations. Also, more risks should be taken in situations with a higher EV for the risky option than for the safe option. The ability to distinguish beneficially between options based on their EV (EV-sensitivity) should increase during adolescence. Here we were specifically interested in whether the association between age and EV-sensitivity would be mediated by working-memory capacity. We investigated 164 participants in a decision-making task, in which the incentive valence (gains, losses) and the EV (risk-advantageous, risk-disadvantageous) were systematically varied. Hence, decisions for the risky option as a function of valence and EV were compared between early (9-11 years), mid (12-14 years) and late adolescence (15-17 years). Results showed (a) higher risk-propensity to avoid losses than to maximize gains, (b) an increase in EV-sensitivity with increasing age, and (c) that this increase was strongly mediated by individual differences in working-memory capacity. These results emphasize the need to consider risk-contexts, as well as cognitive control abilities such as working-memory in theories concerning adolescent development, as they seem to influence the ability to make reasonable decisions in risk-taking situations.

ABSTRACT. People with mental disorders have to overcome multiple barriers to find a way back into full occupational reintegration. With this project we aim to get a better understanding of the processes underlying the stereotype-based employment discrimination applying a cognitive decision making approach. With a hypothetical reintegration hiring scenario with probabilistic inference decisions we show that participants base their response not only on the information about the actual (objective) suitability of the applicants but also systematically discriminate against people with mental disorders compared to people with physical disorders (Study 1), and that individuals are not able to correct for this discrimination even with explicit instructions and monetary incentivization (Study 2). We show that this hiring bias is mediated by the perceived competence of the applicants. Furthermore, we show that the hiring bias varies between different occupations used for the scenario (i.e., architect, manager and police officer) supporting the assumption that the stereotypical perception of a certain job’s requirements is also crucial for employment discrimination. In sum, results suggest that people activate stereotype knowledge associated with mental and physical disorders serving as cues in probabilistic inference decision tasks in the context of employment decisions. These cues result in systematic discrimination even under incentives that make discrimination irrational.

ABSTRACT. The validity of surveys asking for self-reports of sensitive attributes is threatened by socially desirable responding. We propose honesty contracts as a new method for reducing social desirability bias. The method is based on placing an explicit contract with voluntary respondents who are asked to commit themselves to providing honest answers prior to starting the survey. Asking participants to actively conclude an honesty agreement increases attention to honesty norms and avoids the problem of participants skimming or skipping instructions asking to provide truthful responses. Honesty contracts are easy to implement and can be added to any survey. They are a rather soft form of nudging because there are no sanctions that force respondents to conclude or to honor a contract. To investigate whether honesty contracts are nevertheless capable of inducing a commitment to answer honestly, we conducted an experiment. We asked 215 participants to provide self-reports with regard to nine sensitive attributes, as for example, driving a car after consuming alcohol, gossiping about other people, and lying to other people. As expected, respondents in the honesty contract condition were significantly more ready to admit socially undesirably behaviors than respondents in the control condition. Virtually all participants (98%) in the honesty contract condition were ready to commit themselves to providing honest answers. Taken together, our results suggest that honesty contracts may be a promising means to reduce response distortions due to social desirability.

ABSTRACT. Previous findings in cognitive decision-making research indicate that decisions get riskier (i.e. more disadvantageous) if a secondary cognitive task is performed simultaneously. Studies often used visual working-memory tasks that had to be performed in parallel to a (visual) decision-making task. Our research question was whether this dual-task interference is also true for decision making in complex situations (including conflicting short- and long-term risks) in case the parallel task addresses a different sensory modality. The current study tested three groups of participants (N = 115), who performed executive function and working memory tasks as well as a complex risky decision-making paradigm (the Cards and Lottery Task; CLT). Participants performed the CLT either as a single task (n = 39) or a dual task including either an auditory 1-back (n = 38) or auditory 2-back task (n = 38). The results show that, overall, performance in the CLT was not affected by a parallel auditory n-back task (neither 1-back nor 2-back). However, the group variable moderated the effects of general working-memory span on the decision-making performance. Furthermore, significant interactions with decision time occurred. The results indicate that working memory contributes differentially to advantageous decisions depending on whether parallel working-memory resources are required or not. However, the results indicate that a secondary cognitive task does not necessarily have negative effects on decision making in complex situations, at least if it addresses a different sensory modality. In this case, having good working memory and taking more time seem to compensate for additional cognitive demands.

ABSTRACT. Previous research has found individual differences in the use of different fast and frugal heuristics suggesting that individuals have preferences for using versus avoiding each of these strategies. Additionally, it has been established that use of a specific heuristic, the recognition heuristic (RH), is stable across time and stimulus materials suggesting that the preferences individuals have for or against using such strategies are stable. These findings indicate that use of heuristics might be a stable person-specific cognitive style of decision making. Extending this line of research, we investigated whether individuals display the same inclination to the use versus non-use of heuristics in general. In particular, we studied use of two of the most prominent heuristics, namely the RH and the fluency heuristic (FH). We hypothesized that individuals who show higher levels of RH-use will also show higher levels of FH-use. To test this hypothesis, we conducted a classical city-size task consisting of a recognition task and a paired-judgment task, which allows the assessment of both heuristics simultaneously. Participants’ proportions of RH-use and FH-use as well as their relation were measured by a hierarchical extension of a multinomial processing tree model (the r-s-model). In line with our hypothesis, a significant positive correlation between use of both heuristics could be observed, i.e., individuals who applied the RH more often also applied the FH more often. This result is in line with previous findings and serves as further evidence for heuristic use representing a stable person-specific cognitive style of decision making.

ABSTRACT. According to the metacontrol state model, cognitive processing can be affected by metacontrol states that vary between two poles: extreme persistence and extreme flexibility. While the former is characterized by a strong top-down influence of the current action goal, high selectivity and goal maintenance, and strong mutual competition between alternative representations, the latter is characterized by a weak top-down influence of the current action goal, low selectivity and goal maintenance, and weak mutual competition between alternative representations. Interestingly, previous research has suggested that metacontrol states can be affected by several long-term factors that may determine relatively stable biases towards either persistence or flexibility. In the present study we focused on the role of culture in shaping metacontrol biases. To this end, we compared performance of people from four different countries (Germany, Italy, The Netherlands, and China; N=60 in each group) on a task assessing cognitive flexibility (i.e., color-shape task-switching paradigm). Results corroborated the hypothesis that whereas collectivistic cultures are more biased towards flexibility, individualist cultures are more biased towards persistence.

ABSTRACT. Emotional information captures attention due to privileged processing. While this emotional-priority is beneficial in some situations, it comes at the costs of increased distractibility in other situations. Therefore, shielding current goals from emotional disrupt (ED) is essential for adaptive goal-directed behavior. Current research on the interplay between cognition and emotion has shown that for negative stimuli, ED is reduced when participants recruit cognitive control prior to the presentation of an emotional distractor. Following up on this finding, we asked whether this control of ED is valence specific or general. A valence-general account predicts that cognitive control shields against the distracting influence of arousal, irrespective of the valence (positive or negative) of an event. In contrast, a valence-specific account predicts that cognitive control interacts with the specific hedonic value of an emotional event and cognitive control reduces ED from negative stimuli only. To test this, we systematically manipulated the valence and arousal (positive, high arousal; negative, high arousal; and neutral, low arousal) of emotional distractors (pictures) and assessed how control (instigated by the flanker task) interacts with ED. We found that following conflict, ED decreased similarly for negative and positive pictures. Accordingly, results support a valence-general account of cognitive control mechanisms that block off high arousing emotional distractors from interfering with goal-directed actions, irrespective of their valence.

ABSTRACT. Many accounts of human action control assume bindings between stimuli and responses of individual events (e.g., Hommel, Müsseler, Aschersleben, & Prinz, 2001; Logan, 1988; Schmidt, De Houwer, & Rothermund, 2016). That is, the representation of an event can be thought of as a loose network of bindings between its elements. Notably, most human action is hierarchically organized, meaning that a particular action (e.g., typing a word) is both part of a larger scale action (e.g., typing a sentence) and also composed of smaller scale actions (e.g., individual keystrokes). Yet, to date, there is no evidence for a hierarchical organization of bindings. The present experiment (N = 32) bridged that gap and analyzed whether bindings within smaller scale events (stimulus-response bindings in individual responses to stimuli) can exist simultaneously with bindings between these events (bindings between successive individual responses). Supporting the notion of a hierarchical organization of bindings, the results indicate that successive individually executed responses can be bound to each other, while (smaller scale) bindings between elements of each individual response exist. Implications for the role of binding in action control in general are discussed.

ABSTRACT. Initial affective reactions to emotional expressions of others have been shown to critically depend on the group membership of expresser and perceiver. The underlying mechanisms that are responsible for this response divergence are, however, still unclear. The most prominent explanation - the social intention account – holds that affective response tendencies are elicited by others’ intentions: In-group members are imputed to pursue benevolent intentions thereby facilitating concordant responses. Out-group members, however, are imputed to pursue malevolent intentions, thus, triggering discordant responses in an interaction partner. In this talk, we put forward an alternative account that is able to explain the observed response divergence much more parsimoniously. We propose that a combination of group membership and facial displays results in affectively compatible (e.g., positive emotional expressions by positively evaluated persons) or incompatible configurations (e.g., positive expressions by negatively evaluated persons) assuming that incompatible faces represent an affective processing conflict. To test this processing conflict account, participants categorized facial displays of in-/out-group persons. Results revealed a clear performance benefit for affectively compatible over incompatible faces (i.e., a compatibility effect) which is indicative of perceived conflict. Moreover, we observed typical conflict adaptation effects: Incompatible faces in trial N-1 substantially reduced the compatibility effect in trial N, which occurred irrespective of the type of the preceding conflict. More precisely, the fact that an affectively incompatible joyful face triggered control adjustments that facilitated processing of an affectively incompatible fearful/angry face, speaks in favor of the conflict instead of the social intention account.

ABSTRACT. In dual tasking, between-task interference arises when both tasks share similar processing characteristics (compatibility-based BCE) or when withholding a response in the secondary task (T2) conflicts with Task 1 (T1) response execution (no-go BCE). Because both types of BCEs seem to fundamentally differ in the underlying cognitive processes (Durst & Janczyk, 2018), it remains unclear whether a contextual modulation of the compatibility-based BCE can be equally assumed to occur for the nogo-based BCE. To test this, a visual S1 was randomly presented at upper or lower locations followed by an auditory S2 (100 vs. 200ms SOA). One location was associated with 80% T2 response execution (go-R2) and the other with 80% not responding in T2 (no-go-R2). Results showed overall slowed R1 when R2 was a nogo-response, demonstrating a no-go BCE. This no-go BCE was contextually modulated: It was present at the context of mostly go-R2s, where withholding R2 considerably slowed R1 execution. At the same time, the no-go BCE was absent at the context of mostly no-go R2s. Frequently withholding R2 at this location, however, did not facilitate R1 execution. Instead, the location of the mostly no-go R2s seemed to generally induce (inhibitory) costs on R1 execution irrespective of R2 response type (go vs. no-go). This modulation was not present at long SOA. Here, participants (strategically) delayed R1, which diminished the predictive impact of the context but maintained an overall no-go BCE. Together, these results demonstrate a contextual modulation of motor-based between-task interference in dual tasks.

ABSTRACT. Sense of agency (SoA) is the sense of having control over one's own actions and through them events in the outside world. People estimate their SoA by integrating different agency cues. In the present study, we examined whether the use of congruency between action and effect, affective valence of the effect, and temporal relation between action and effect as agency cues differs between Asian (Mongolia) and Western (Austria) cultures. 61 Mongolians and 70 Austrians participated. In a learning phase, participants learned to associate actions (keypresses) with positive and negative action effects. In a test phase, participants performed the same keypresses. After different intervals positive and negative action effects, which were either congruent or incongruent with the previously acquired action-effect associations, were presented. In each trial, participants were asked to rate how likely the action effect was caused by themselves or by the computer (authorship rating). Higher authorship ratings for congruent than incongruent action effects indicated that action-effect congruency strengthens SoA. Higher authorship ratings for positive than negative action effects indicated that emotional valence modulates SoA. Authorship ratings decreased with increasing interval in Austrians but not in Mongolians. For Mongolians, the temporal chronology of events might be less important when inferring causality. Therefore, information regarding the temporal occurrence of the effect might not be used as agency cue. In conclusion, some agency cues might be similarly used in different cultures but the use of other cues might be culture-dependent.

ABSTRACT. Chronic low blood pressure (hypotension) is typically associated with symptoms including fatigue, mood disturbance, dizziness or cold limbs. In additions, a number of studies demonstrated hypotension-related cognitive impairments, particularly in the domains of attention and memory. Deficiencies in cerebral blood flow regulation may contribute to these deficits. This study investigated cerebral blood flow modulations during proactive control in hypotension. Proactive control refers to cognitive processes during anticipation of a behaviourally relevant event that allow optimization of readiness to react. Using functional transcranial Doppler sonography, bilateral blood flow velocities in the middle cerebral arteries (MCA) were recorded in 40 hypotensive and 40 normotensive participants during a precued Stroop task. The task included 18 congruent and 18 incongruent trials, which were presented 5 s after an acoustic cue. The MCA supply cerebral structures as the dorsolateral prefrontal and inferior parietal cortices, which are relevant in preparatory cognitive processing. Hypotensive participants exhibited smaller bilateral blood flow increases during response preparation and longer response time. The group differences in blood flow and response time did not vary by executive function load, i.e. congruent vs. incongruent trials. Over the total sample, the flow increase correlated negatively with response time in incongruent trials. The findings indicate reduced cerebral blood flow adjustment during both the basic and more complex requirements of proactive control in hypotension. They also suggest a general deficit in attentional function and processing speed due to low blood pressure and cerebral hemodynamic dysregulations rather than particular impairments in executive functions.

ABSTRACT. Performing two tasks simultaneously involves the coordination of their processing. Task coordination is particularly required in dual-task situations with varying order of the component tasks. When task order switches between subsequent trials, task-order coordination leads to order switch costs in comparison to task order repetitions (i.e., worse performance in trials associated with an order switch compared to an order repetition). However, the adaptive characteristics of task-coordination processes and order switch costs are underspecified so far. For example, studies on other control processes have shown that these processes can be modulated in response to changes in task demands. The present study investigated therefore whether task-order coordination processes are modulated by the previous experience of a task-order switch. To investigate these costs in a dual-task situation with two sensorimotor tasks with variable task-order, we analysed reaction times in trials with task-order switches and trials with task-order repetitions following task-order switches and repetitions in the preceding trial. Order switch costs were reduced in trials following task-order switches compared to task-order repetitions; resembling the Gratton effect commonly observed in conflict adaptation paradigms. We discuss the present results in the context of cognitive control theories.

ABSTRACT. We investigate the reciprocity of task choice and task performance in multitasking when participants can freely decide which task to perform on each trial. It is a central finding that switching between tasks leads to switch costs. Thus, in voluntary task switching experiments the switch rates were often low. In the self-organized task switching paradigm, a novel variant of voluntary task switching, the likelihood of tasks switch is increased because the stimulus needed to repeat the previously executed task, is delayed. In each trial, participants freely decide whether to categorize a letter stimulus as vowel or consonant or a number stimulus as even or odd. The stimulus, needed for task switch occurs immediately, while the stimulus associated with repeated task appears delayed. This delay increases with each repetition until the participant decides to switch. Previous research has shown that this delayed presentation of the repetition stimulus leads to increased switch rates. Moreover, switch costs and waiting time for repetition stimulus were rather similar. In the present study, we aim to optimize the novel self-organized task switching paradigm. We elaborated whether delay increment per repetition and switch costs impact on task choice to asses which combination of delay increment and switch costs maximizes participants’ attempt to tradeoff waiting time and switch costs. Correlational analyses indicate relations between individual switch costs and individual switch rates across participants. Moreover, we identified conditions where participants could trade their switch costs and waiting time for the repetition stimulus most efficiently. 245

ABSTRACT. In our modern world, we are constantly faced by an endless array of stimuli, all put different demands on our attentional system and, moreover, require us to permanently switch between these demands. While previous task-switching paradigms mostly focused on the rapid switching between small sets of simple tasks, or different modalities, switching between different attentional demands (such as selective and divided attention) is somehow neglected. Therefore, we developed the Switching Attentional Demands (SwAD) paradigm to investigate switching costs between different attentional demands. The study at hand addresses relationships between the SwAD paradigm, the UPPS impulsive behavior scale, an oddball task, a classic dual-task paradigm, as well as a visual digit span task. In total, 92 people [M=21.59, SD=3.66; 75 women] completed the study. Results showed – among others – that switching between attentional demands leads to an increase in RTs of selective attention (t(91)=78.256, p<.001). In contrast, RTs in the divided attention task decreased under switching conditions, compared to the performance of solely divided attention tasks (t(91) = 72.122, p < .001). Significant correlations were found between RTs of selective attention and the oddball task (r(87)=.536, p<.001), as well as between RTs of divided attention and a classic dual-task paradigm (r(75)=.444, p<.001). Switching between attentional demands had a differential effect on selective and divided attention. Future studies should implement further demands such as vigilance or sustained attention within the SwAD paradigm, but also supplement behavioral findings by investigating neurophysiological correlates.

ABSTRACT. Procedural mistakes committed by the researchers during their activity throughout research data management can deteriorate both the credibility and efficiency of the research project. In our exploratory study, we investigated the frequency and seriousness of mistakes that researchers commit during data management. We surveyed 484 psychology researchers about the most frequent and most serious mistakes that they committed in the last 5 years. We collected the contact information of our sample from articles published between 2010 and 2018 in 153 psychology journals. The results showed that 72% of research teams discover data-management issues in their projects, for more than half of them leading to major project delays or money loss. By exploring the most frequent and serious mistakes we can improve the efficiency and credibility of psychological research in the long run by linking existing solutions to the problems or to show where solution-development is most needed.

ABSTRACT. Our daily experiences are usually made up of multistep sequences of actions and sensory events. Events trigger actions, which in turn evoke ensuing events and further actions. Previous studies on dyads consisting of an action and an event revealed a perceived shortening of the interval between causally linked actions and events, i.e., temporal binding. That is, the perceived points in time of actions and events are shifted towards one another. What happens to this perceived shortening when actions and events are presented in multistep sequences? Previous research examining the perceived timing of events in sequences suggests that action-event sequences break down into dyads which are temporally grouped together. It is not clear, however, whether actions are equally drawn towards the event in multistep sequences. This study examined temporal binding in event-action-event and action-event-action triads. Therefore, participants (N=36) had to judge the perceived timing of actions in six different conditions. An action-only baseline condition was compared to the experimental conditions comprising either an action-event dyad or triads of actions and events. Actions in dyads were temporally bound to both preceding and ensuing events. However, the perceived action-timing in action-event-action sequences did not differ from the baseline condition. Contrarily, the action in event-action-event sequences was perceived to occur earlier than in the baseline condition. Consequently, the results suggest that while actions and events are temporally drawn together in isolated dyads this pattern is more complex in multistep sequences of actions and events.

ABSTRACT. Background and aims: The pathological use of social media applications is discussed as specific type of Internet-use disorders, sharing similarities in behaviors and underlying processes with Internet-gaming disorder. Based on current theoretical assumptions and research on Internet-gaming disorder as well as substance-use disorders, it is assumed that the confrontation with addiction-related cues also has an effect on Internet-related decision-making behavior as well as on executive functions. Methods: The current experimental study used a between-subjects design (N=356) with two groups performing a working memory task (either with neutral or social-media-related cues) in parallel to a decision-making task (Game of Dice Task; GDT). We furthermore investigated associations between task performance and tendencies towards an addictive use of social media (assessed by a modified version of the short Internet Addiction Test). Results: In the social-media-related version of the working-memory task, participants made more mistakes compared to the neutral version. Additionally, lower performance in the social-media-related working-memory task was associated with an addictive use of social media. GDT performance was not different across the two groups. Conclusions: The results are consistent with theoretical models as well as empirical results in addiction research, which emphasize that addiction-related cues affect attention and working-memory capacity. Reductions in the addiction-related working memory task are associated with the urge to use these applications, but not with decision-making behavior. Future studies should investigate the relevance of further executive functions such as inhibitory control when being confronted with specific cues in the context of addictive social media use.

ABSTRACT. In the context of effort-based decision making, it is assumed that humans generally tend to avoid effort. However, effort is a vague concept and may be experienced differently between individuals. In a recent study, we have already shown that some participants deliberately invested more effort than others as indicated by more frequent switches to the more difficult of two tasks (especially so with an increasing reward prospect), even though the potential reward increase was not contingent on task choice but on performance only (Jurczyk et al., in press). In order to investigate whether the subjective cost of effort may explain this seemingly irrational behavior, we used the same paradigm of voluntary task switching between an easy and a difficult task and reward cues of changing magnitude (N=64). After this task, we measured each individual’s subjective effort cost: The effort discounting paradigm (Westbrook et al., 2013) requires participants to make a series of choices to indicate their preference between re-doing a difficult-task block for a certain amount of money or an easy-task block for less. In order to pinpoint the subjective cost of the difficult block (versus the easy block) for each participant, the difference between the two reward offers adapts based on participants’ prior choices, approaching a subjective indifference point. Results show the predicted correlation between the subjective value of effort and individual’s voluntary choice rate of the difficult task, particularly in a situation of increasing reward prospect. Implications for the study of effort-based decision making will be discussed.

ABSTRACT. Content variability was previously suggested to promote stronger learning effects in cognitive training whereas less variability incurred transfer costs (Sabah et al., 2018). Here, we expanded these finding by additionally examining the role of learners' control in short-term task switching training by comparing voluntary task switching to a yoked control forced task switching condition. To this end, three conditions were compared: a (1) Voluntary and a (2) forced task switching group with changing tasks in every block and (3) a voluntary task switching group with the same two tasks in all training blocks. The experiment consisted of a baseline block, seven training blocks (learning phase), followed by one transfer block. Although no additive benefits for learners’ control beyond varied content training was observed, granting participants control over learning seemed to counteract the entailed costs in the third group when transferred to untrained task switching context. Taken together, the current results provides further evidence to the beneficial impact of variability on training outcomes. In addition, they encourage the reconsideration of motivational differences as moderating rather than confounding factors in cognitive training research.

ABSTRACT. Reacting to a stimulus can lead to benefits or interference in a succeeding trial, depending on response and feature repetitions: If responses and features only partially repeat, interference occurs, leading to slower reaction times and higher error rates. However, if responses and features fully repeat, typically benefits are observed. One explanation for this pattern is feature-response binding: Response and stimulus features are bound together and can be retrieved, when repeated. This leads to benefits for full repetition, but interference for partial repetition. Even irrelevant information, i.e. distractors, can be bound to a response (distractor-response binding). According to the literature, such binding mechanisms are ubiquitous and can possibly explain a large amount of actions. We investigated, if these binding effects are effector-specific, specifically, if binding effects are present when giving saccadic responses. In an eye-tracking study (N=27), participants had to discriminate appearing target letters by looking to one of two locations. The appearing target letter was framed by a distractor. We found no evidence for distractor-response binding. These results suggest that eye-movements follow other action control processes than manual movements do.

ABSTRACT. Preceding a difficult action, e.g., a penalty kick, time appears to slow down. Underlying this subjective dilation of time might be the motivational significance of the event and the resulting expenditure of effort prior to the action. On a neural level, this is related to the phasic release of norepinephrine (NE), which has recently been shown to contribute to subjective time dilation. In order to test the assumption that preparation-related time dilation is caused by a phasic NE release, we investigated how indices of NE release, the P3b amplitude and pupil dilation, as well as time estimation are modulated by the cued difficulty of a subsequent mouse-tracking task. We presented participants stimuli of varying duration that cued the difficulty (easy, hard) of the immediately following task. This task involved moving a cursor into a target area within a time limit. The target area was either large (easy) or small and accompanied by distractors (hard). Following this task, participants had to estimate the cue duration relative to two previously established references. As hypothesized, cues indicating a difficult task were not only perceived to be longer than cues indicating an easy task, they also caused a more pronounced P3b amplitude and a larger pupil dilation response. Replicating prior results, we further found that the P3b amplitude predicted subsequent time estimates. In sum, these results indicate that during the preparation of a difficult task the recruitment of cognitive effort, which manifests in the increased release of NE, causes subjective time dilation.

ABSTRACT. According to load theory, task-irrelevant stimuli are processed automatically under conditions of low perceptual load. This notion was challenged by the discovery that, even under low perceptual load, thought-to-be irrelevant stimulus features are not processed. In the so-called concentric circle task (Eitam, B., Yeshurun, Y., & Hassan, K., Journal of Experimental Psychology: Human Perception and Performance, 39, 611–615, 2013), participants saw two differently colored concentric circles – a ring surrounding a disk. Participants were instructed to concentrate on one of the two objects. After a brief exposure to the stimuli, in an unannounced memory task participants had to identify the color of either the previously focused on stimulus (congruent condition) or the color of the previously irrelevant stimulus (incongruent condition). Memory for the previously irrelevant stimulus color was significantly worse. In our present one-trial experiment, we aimed to conceptually replicate the original study while addressing a variety of caveats of the original study and measuring eye-movements in order to implement a more sensitive measure of residual memory (i.e., priming). Although we only partially replicated the findings of the original study, a meta-analysis of reported irrelevance induced blindness effects lends credence the existence of this effect.

ABSTRACT. Visual complexity is a construct of relevance in many different domains, from basic research to human-computer interaction or the perception of art. However, despite its relevance, visual complexity is still not understood comprehensively. Within a novel approach, we experimentally investigated the influence of core dimensions of visual complexity on subjective ratings in order to contribute to a better understanding of the construct itself. Moreover, we employed eye tracking in order to gain deeper insights into the perception of complexity. We hypothesized that both the number of elements as an aspect of the quantitative dimension as well as symmetry as a facet of the structural dimension strongly affect the participants’ subjective perception of visual complexity as well as their gaze behaviour. For that purpose, we constructed a controlled picture set which allowed for the manipulation of the above-mentioned dimensions. These stimuli were presented within a laboratory eye tracking study with 33 participants. Statistical analyses revealed significant main effects of both number of elements and symmetry on ratings of visual complexity. Moreover, effects of the experimental manipulation could also be found within the ocular parameters number of fixations, scanpath length and spatial density. Our results support a two factor model of visual complexity as reported by other researchers (Chipman, 1977; Gartus & Leder, 2017; Ichikawa, 1985), while contributing to a better understanding of the cognitive processing of visual complexity.

ABSTRACT. The ability to evaluate the plausibility of informal arguments is an important competence for making sense of scientific texts across different domains. To investigate the cognitive processes involved in evaluating informal scientific arguments, we used eye-tracking measures and related these measures to the accuracy of the plausibility judgements provided during reading scientific texts. Indicators of strategic processes (e.g., regressions and longer reading times), which reflect an elaborative processing of cognitive conflicts, were predicted to be positively related to accurate plausibility judgments for implausible (invalid) but not for plausible (valid) arguments. The sample (60 university students) consisted of mostly female participants (83.3%) with an average age of 25.7 years. Participants were asked to judge the plausibility of 22 informal arguments embedded in two coherent scientific texts about psychological topics. Subsequently, the students assigned the sentences marked as implausible to one of several categories of reasoning errors. To vary the arguments’ plausibility, ten of the 22 arguments contained typical reasoning errors. Linear mixed models revealed differential relations between the accuracy of plausibility judgements and eye tracking indicators for implausible and plausible arguments. Implausible arguments that were judged accurately were related to more regressions and longer reading times, indicating a more elaborated and integrative information processing. As hypothesized, this was not found for plausible arguments. These arguments do not require elaborated processing strategies because less cognitive conflicts occur. These findings indicate that different processing strategies are employed for evaluating scientific arguments, depending on characteristics of the arguments and the individual reader's competence.

ABSTRACT. Visual processing tasks, such as reading, involve the perception of visual symbols with series of fixations, their short-term storage and decoding. For fluent processing in such a task, the ability to perceive multiple symbols with single fixations is crucial. As of yet, it is an unresolved question how this ability is influenced by memory load. I expect that limited attentional resources are used to ensure a high perceptual capacity of fixations. Hence, the amount of information that can be perceived with a single fixation should decrease with increasing memory load. I employed a novel combination of the complex span paradigm and eye tracking to test this assumption. Music students (n=75) were asked to memorize one note and then play a simple melody at first sight on a piano. After twelve repetitions of this procedure, they were asked to recall the memorized notes in correct order. Eye movements during the performance of the melodies were tracked. While the distance of saccades was unaffected by memory load, the number of fixations used to read the melodies increased with each additional note that had to be held in memory. When more notes were stored in immediate memory, fewer attentional resources were available for visual processing and the amount of information that was perceived with a single fixation decreased. These findings show that the combination of complex span tasks and tracking of eye movements are the ideal research method to analyze the influence of memory load on eye movements during visual processing.

ABSTRACT. Performance tests to estimate someone’s ability can be hampered by ceiling and floor effects, i.e., if certain tasks are too easy or too hard for most of the participants. Even differences in reaction times might be uninformative in extreme cases. Contrary, gaze-fixation-based data seem to be a promising alternative to deliver further insights in cognitive processes that might be related to abilities in a specific domain. This hypothesis was examined with N = 48 participants in the context of spatial thinking utilizing the R-Cube-Vis test consisting of six distinct difficulty levels, whereby the easiest and the most difficult level usually produce ceiling and floor effects, respectively, in a student population. Participants performed the test while tracking their gaze movements. Afterwards, 2-states Hidden Markov Models (HMMs) were computed, one for each participants and each level intended to reflect cognitive processes for this kind of tasks. A certain parameter of the HMM, which might indicate an uncertainty about the structure of the item, was able to significantly predict the test performance of the whole test additional to the accuracy and reaction times of the respective level, but only in the easiest level. Although, similar effects for the most difficult level could not be found, the results strongly support the potential of gaze fixation based measures to indicate cognitive processes to deliver further insight in a participant’s ability that goes beyond the information provided by accuracy and reaction times.

ABSTRACT. A robust finding in eye research shows that blinks increase their rate if a person indulges in a conversation compared to quiet rest. Various factors, such as social demands, emotional engagement, cognitive load and motor actions during speaking, were suggested to explain this increase. One main criticism of the motor based explanation comes from the finding that chewing a gum do not result in an equally increased blink rate compared to talking. Stimulation of facial motor neurons has been shown to innervate both the facial eye muscles as well as muscles around the lips. We therefore hypothesize that we must differentiate between muscle groups in order to assess the influence of motor components on the blink rate. Analyzing eye tracking data of 27 subjects sitting alone in a room (without social interaction), we replicate the significant difference in blink rate between talking and resting. A comparison between different motor output conditions further showed that while talking without sound and eating a lollipop led to a significant increase in blink rate, chewing gum did not. Neither auditory nor cognitive task modulations showed an influence on the blink rate, however, interaction effects might be present. Our findings clearly suggest a relationship between blinking and motor output but indicate a specific role of lip movements. Such purely motor related influence on blink rate advises caution when using blinks as neurological indicators during patient interviews. Further research needs to investigate if cognitive and social factors have an independent influence on the blink rate.

ABSTRACT. Sensory signals provide ambiguous information about the environment. If a stimulus evokes several interpretations ("percepts") that alternate in awareness, this is referred to as "multistability". For visual multistability, it has been demonstrated that the pupil dilates when the percept changes ("perceptual switch"). We ask whether the pupil reacts similarly to perceptual switches in auditory multistability. We presented a sequence of two alternating sounds ('ABABAB...') in three conditions. In "rivalry", participants reported their dominating percept by pressing one of four buttons, referring to four reportable percepts: perception of a coherent 'AB' sequence ("integrated"), or perception of separate 'A' and 'B' sound streams ("segregated") with either 'A' or 'B' or both streams in the foreground. In "replay", the perceptual switches of a preceding rivalry block were simulated by physical stimulus changes, and participants reported analogously to rivalry. In the "random" condition, participants listened to the stimulus of the rivalry condition and pressed the four buttons randomly. We found the pupil to dilate around perceptual switches. Dilation started prior to the report of the switch and peaked 750ms after. Dilation in response to stimulus changes in replay was larger than to perceptual switches in rivalry. In the random condition, participants exhibited a qualitatively different button-press pattern, rendering this condition incomparable to rivalry and replay. The observed pupil response in auditory multistability shows similarities to previously reported effects in vision. This suggests common principles to be involved in multistability in both sensory modalities.

ABSTRACT. Modernization of assessment and enhancement of human performance capabilities among commercial and civilian aircrew is becoming increasing important to sustain the safety of aviation operations across international airspace. When selecting pilots, mainly cognitive tests are used as this has shown to be a good indicator of pilot success across a range of studies. Non-cognitive aptitudes such as stress tolerance are rarely assessed, but have shown to be relevant for high performance as well. Our main research question is whether pilots with high non-cognitive aptitudes show higher performance than pilots with low non-cognitive aptitudes. We plan to conduct an experiment in a flight simulator with 30 pilots or advanced trainees. The study will be conducted in a multi-level assessment approach involving psychophysiological data (eye-tracking), subjective data, and performance data. Preceding the simulation, non-cognitive aptitudes such as stress tolerance and emotional regulation will be assessed. The participants will be divided into two groups before the main experiment starts: (1) Lower non-cognitive aptitudes, (2) Higher non-cognitive aptitudes. Each pilot will have to fly different maneuvers under different flight conditions. Levels of stress will be induced, divided into two levels: 1) low (e.g. instruments on); 2) high (e.g. instruments off). Participants will be instructed to maintain heading in a sensory-deprivated environment. We suppose better performance for pilots with high non-cognitive aptitudes than for pilots with low non-cognitive aptitudes; the difference should be the strongest under high stress induction. The experiment will be conducted between April and June 2019.

ABSTRACT. Deriving a diagnosis in medical as well as technical settings is a complex task that requires expertise and involves problem solving. Recent research could show that the transfer of an expert’s eye movement can have beneficial effects, both during remote collaboration and as modelling examples during training. Interestingly, these effects did not appear reliably. We propose the method of visualisation as a key factor for successfully transferring eye movements. Consequently, our research examines the question, which characteristics of visualisation methods can improve the perceptibility and utility of transferred eye movements. In an experiment (N=50) we contrasted three different visualisation methods, which differed regarding the amount of information displayed, with a control condition. The visualisations were presented either at original or at half speed, resulting in a 2x4 mixed design. The stimulus material consisted of screen recordings of experts solving the Tower of London task. The results show significantly higher workload ratings for all visualisation methods compared to the control group. More complex visualisations showed better ratings regarding ease to follow while their perceived overall usability depended on the playback speed. Participants remembered details of the solution process equally well in all conditions but performed significantly better at the Tower of London task after watching a gaze visualisation of lower complexity compared to the control condition. We conclude that potential benefits of transferred eye movements not only depend on both speed and amount of displayed information, but also on the measure chosen for evaluation.

ABSTRACT. Ideomotor theories suggest that goal-directed actions are selected based on learned contingencies between actions and their effects (e.g., Elsner & Hommel, 2001). These action-effect associations let us anticipate the future consequences of our actions which is also evident in anticipatory saccades towards the location of future effects (Pfeuffer, Kiesel, & Huestegge, 2016). These anticipatory saccades reflect a proactive effect monitoring process which prepares a later comparison of expected and actual effect. Interestingly, anticipatory saccades are initiated earlier when the interval between action and effect is shorter. Here, we questioned how such anticipatory saccades are affected by violations in participants´ temporal effect expectations. Left/right keypresses generated effects (colored circles on the left/right side) that predictably appeared after a short or long action-effect interval (75% frequent action-effect interval/25% infrequent action-effect interval; N = 32). Each keypress consistently produced spatially action-effect compatible/incompatible effects per half of the experiment. First, we replicated the finding that anticipatory saccades (as well as manual responses) occurred earlier for short rather than long action-effect intervals. Anticipatory saccades towards future action effect locations were more frequent and occurred earlier following trials with infrequent rather than frequent action-effect intervals. A similar pattern emerged for manual responses. Moreover, when action-effect intervals were longer than expected, expectancy violations also showed in fewer but earlier anticipatory saccades towards future effect locations in those trials themselves. Thus, we conclude that proactive effect monitoring processes can be adapted based on violations of temporal effect expectations.

ABSTRACT. In addition to symptoms including fatigue, dizziness, reduced drive or mood disturbance, individuals with chronic low blood pressure (hypotension) frequently report cognitive impairments. While a number of studies confirmed reduced performance in attention and memory, not much is known about hypotension-related deficits in executive functions. This study investigated cerebral blood flow modulations in hypotension during a precued antisaccade/prosaccade task requiring the executive function of proactive inhibition in addition to preparatory attention. Using functional transcranial Doppler sonography, bilateral blood flow velocities in the middle cerebral arteries were recorded in 39 hypotensive and 40 normotensive participants. In the task, a stimulus appeared left or right of a fixation point 5 s after a cuing stimulus; subjects had to move their gaze to the mirror image position of the stimulus (antisaccade) or towards it (prosaccade control condition). Video-based eye-tracking was used for ocular recording. A right dominant MCA blood flow increase arose during task preparation, which was smaller in hypotensive than normotensive participants. Moreover, hypotensive participants exhibited lower peak velocity of the saccadic response. The extent of the reductions in blood flow and task performance in hypotension did not differ between antisaccade and prosaccade trials. The smaller MCA flow increase may reflect reduced activity in the dorsolateral prefrontal and inferior parietal cortices during proactive inhibition and preparatory attention in hypotension. Given that group differences in blood flow and performance arose independent of task complexity and executive function load, hypotension may be characterized by basic attentional impairments rather than particular executive function deficits.

ABSTRACT. The concurrent execution of cognitive and postural tasks is associated with an increased risk of falls in older adults. Biological aging of the cognitive and postural control system appears to be responsible for increased cognitive-motor interference effects. We examined neural and behavioral markers of motor-cognitive dual-task performance in young and old adults performing spatial 1-back working memory single- and dual-tasks during semi-tandem stance. We used EEG to test for age-related modulations in the frequency domain related to cognitive-postural task load. Twenty-eight young and thirty old adults performed a postural single-task, a cognitive-postural dual-task and a cognitive dual-task with postural demands (triple-task). Postural sway was recorded in semi-stance on an unstable surface placed on a force-plate during task performance. Neural activation was recorded using a 64-channel mobile EEG system. Our findings revealed impaired cognitive dual-task performance in old participants, as indicated by significantly lower cognitive performance at triple task demands. Further, old adults showed significantly larger postural sway, especially at cognitive-postural task conditions. Old participants also showed significantly lower theta- and alpha-band activity and greater beta-band activity, most pronounced at Fz electrode. Moreover, we observed higher theta- and lower alpha-band activity with increasing task difficulty. These effects were greater in young adults for theta- and alpha-band activity, suggesting an age-related cognitive decline. The results suggest that old compared to young adults were able to recruit additional neural resources at low task demands but not at high task demands, reflecting a compensatory mechanism in order to cope with an age-related cognitive decline.

ABSTRACT. In visual search tasks, participants have to locate and identify a target stimulus within a set of competing distractors. Errors can happen easily in such displays due to interference from the more salient distractors. Neural processing of these kinds or errors are often investigated by means of speeded choice tasks that require a motoric response (e.g. a button press). They typically feature two response-locked neural correlates of error processing, the Ne/ERN and the Pe. The conflict monitoring theory suggests that the Ne/ERN emerges due to a motor conflict between competing response tendencies. However, in the case of visual search, errors can occur that are purely attention-based and therefore independent of any motor response. Here, we aim to investigate the occurrence of neural correlates of error processing after such covert, non-motor errors. To this end, we collected and analysed EEG data from 30 participants completing a multi-frame visual search task, in which an overt motor response was required only after completion of all display frames. This temporally detaches any motor response from the time point of error occurrence. Our analyses indicate that error processing is triggered after erroneous target selection alone, as the Ne/ERN emerges even though the target selection is accomplished without any overt motor response. This demonstrates that non-motor errors are associated with similar mechanisms of error processing as motor errors.

ABSTRACT. Behavioural studies have shown connection between impulsivity and verbal performance. The accuracy on the verbal tasks correlated negatively with the level of impulsivity. Our aim was to investigate the neural background of the previously found correlation between verbal performance and impulsivity. Our hypothesis was that impulsivity will have an effect on the activation of Broca-related language networks. The sample consisted of ninety healthy, right handed university students (45 males), who participated in a verbal fluency task in the fMRI, and they filled out two self-reported questionnaires targeting impulsivity (BIS-BAS, BIS11). Task-related activations and deactivations were assessed using block design fMRI in a phonetic verbal fluency task. The paradigm included seven cycles of 30-second long altering rest and internal word generating task. During the active condition, participants had to silently generate different words starting with a particular letter (S, K, E, T, L, A, N), without any movements. During the rest periods, they had to relax. The measurements were performed on a 3T Magnetom Tim TRIO, human whole-body MRI scanner (Siemens AG, Erlangen, Germany) with a 12-channel head coil. To assess task-related networks we used model-free independent component analysis (ICA) on the fMRI data. The results showed that the activation measured in the Broca area confirm our hypothesis. From the 21 task-related meaningful components we found, 4 correlated positively with the behavioral activation system (BAS) and 7 with the motor impulsivity factor of BIS11, therefore the people with higher level of impulsivity had a greater activation in the language networks.

ABSTRACT. Whilst it has long been known that the cerebellum is crucial in coordinating movements, more recent studies have also implicated the cerebellum in executive functions such as dual-tasking. However, it is still unclear which precise mechanisms of motor-cognitive dual-tasking are supported by the cerebellum. We hypothesized that the cerebellum would integrate the two single tasks into a dual-task network. We created a paradigm to assess dual-tasking abilities using a continuous motor task in addition to a cognitive visual attention task based on the “theory of visual attention” (TVA). 26 patients with isolated cerebellar lesions and 26 healthy matched controls were tested using this paradigm. fMRI data was also collected. Both groups performed similarly in the TVA-based task in both the single and dual-task conditions, with no significant differences between the groups in their visual threshold, visual processing rate, and the visual short-term memory storage capacity. However, despite showing no residual deficits in fine motor skills, the patient group showed much greater difficulties in performing the motor task in the dual-task condition, although not in the single-task condition. Moreover, resting state connectivity data indicated a correlation between the performance in the motor task in the dual-tasking condition and the connectivity between the cerebellum and the ventral attention network stream. These data suggest that the cerebellum plays an important role in the integration of both tasks during motor-cognitive dual-tasking. Moreover, in cognitively demanding situations such as motor-cognitive dual-tasking, a cerebellar lesion may impede the distribution of attentional resources.

ABSTRACT. The question if the processing of emotional content in the absence of awareness is possible cannot yet been conclusively answered, since heterogeneous methodological procedures lead to heterogeneous results. Here, an attentional blink (AB) paradigm was used during functional magnetic resonance imaging (fMRI) to render emotional faces invisible and investigate the neural correlates of non-conscious processing under limited attentional resources. AB is a rapid serial visual presentation (RSVP) paradigm in which participants are, usually unable to report a second (T2) of two targets if it occurs 200-500ms after the first one (T1). Since T2 fails to reach visual awareness despite considerable visual processing we expected, based on the prominent theories in this field, at least subcortical activity for unseen emotional faces. Before the main experiment was carried out, all participants (n = 40) performed an AB-training session followed by an individual contrast calibration (CC). In the CC the AB-task was performed and contrast was calibrated for faces to achieve a comparable trial number of seen and unseen emotional trials. Beside the expected activity for seen emotional faces, we observed neuronal activity for emotional compared with neutral faces regardless of visibility. This is in line with our hypothesis, that processing of non-conscious emotional stimuli is possible when unawareness is caused by a lack of attentional resources.

ABSTRACT. Cognitive control processes are important in order to pursue goal-directed action in conflicting situations. In recordings of electrical brain activity, oscillatory theta activity as well as the N2 and P3 event-related potential components have been identified as electrophysiological markers of conflict-related processing. The aim of this study was to compare both electrophysiological markers of conflict processing in two different conflict tasks. The Go/NoGo task was used to induce response inhibition, whereas the Eriksen flanker task was administered to probe interference control. While the subjects executed the tasks, an electroencephalogram was recorded from 64 scalp electrodes. The sample consisted of 18 women and 9 men aged between 19 to 29 years. In the analysis of event-related potentials, the N2 was modulated by conflicting trials in both tasks (NoGo trials, incongruent trials in the flanker task). Only in the Go/NoGo task a P3 effect (NoGo P3) was observed, suggesting that this ERP component is specifically associated with motor control. Time-frequency analysis showed increased theta power in conflict situations in both tasks. The results of our study suggest that brain processes given rise to the N2 component as well as to oscillatory theta activity support conflict resolution in both response inhibition and interference control tasks. The P3 ERP component, in contrast, specifically reflects motor control during response inhibition in the Go/NoGo task.

ABSTRACT. Although a substantial body of research has investigated the effects of aerobic exercise on cognitive performance, few have monitored exercise-concurrent cognitive processes via electroencephalography (EEG) and fewer still using an event-related potential (ERP) approach. As such, little is known regarding how the temporal dynamics of cognitive processing are influenced during aerobic activity. Here, we aimed to elucidate the effect various modalities of aerobic exercise might have on the temporal dynamics of concurrent visual working memory (VWM) performance. Participants (n = 18) performed a VWM retro-cue task at rest and during aerobic exercise across two postural modalities: seated (using a stationary bicycle) and standing upright (using a treadmill). Three consecutive phases of the VWM processing pipeline were assessed via lateralized ERPs – access to VWM representations (CDA), response selection (sLRP), and response execution (rLRP). Aerobic exercise and upright posture were found to significantly facilitate VWM performance; reaction times (RTs) were expediated during exercise, while both RTs and error rates decreased during upright posture. Further, analysis of ERL waves isolated the observed speed of processing facilitations to a finite temporospatial stage of the cognitive processing pipeline, between the phases of accessing VWM representations (CDA) and response selection (sLRP). Our findings hold implications not only for understanding the influence of aerobic exercise on VWM, but also for contemporary models of VWM that are built almost exclusively on data recorded during seated rest.

ABSTRACT. Goal-directed behavior requires constant monitoring and rapid evaluation of ongoing actions. The error-related negativity (ERN) has been studied extensively as a neurocognitive electrophysiological correlate of action monitoring, yet little is known about links to the affective significance of errors. The current study aims to elucidate how action monitoring relates to affective processing by investigating the relationship between the ERN and activity of the autonomic nervous system and by examining whether action monitoring processes involve an affective evaluation of performed actions. In our first experiment, EEG and skin conductance response (SCR) were recorded while participants (N = 29) performed a go/no-go task. Results indicated that the ERN amplitude predicted a heightened SCR to incorrect responses on a single-trial level. In our second experiment, we assessed behavioral performance during a go/no-go task, which was combined with a word categorization task. After each response to the go/no-go stimuli, participants (N = 30) categorized an affective word as either positive or negative. Extending previous findings on affective priming, incorrect responses were followed by faster categorization of negative words, while correct responses were followed by faster categorization of positive words, suggesting that responses were automatically appraised along a negative-positive dimension. Integrating our results into current theoretical frameworks, the relationship between SCR and ERN supports the notion of the ERN as an alarm signal, reflecting the affective significance of errors. A relation between action monitoring and affective processing is substantiated by results of experiment 2, indicating that an affective value is rapidly assigned to performed actions.

ABSTRACT. Words encoded in the context of a survival scenario are better remembered, an effect that is modulated by word imageability or concreteness. To investigate the contribution of different retrieval modes to this “survival processing effect”, we examined measures of familiarity- and recollection-based retrieval using receiver operating characteristic (ROC) parameters and brain activity (event related potentials; ERPs). Thirty-six participants encoded high- and low-imageability words by rating their relevance either to a survival scenario or to a moving (control) scenario. Performance on a subsequent recognition test replicated the survival processing effect – better memory performance for the survival group – only for words that were high in imageability. The ROC analysis revealed that high-imageability words were associated with greater recollection than low-imageability words, and that this effect was increased in the survival group. ERPs elicited during the recognition test showed a late old/new effect 500-800 ms after word onset, which is typically associated with recollection-based retrieval, for both groups and word types. However, only for the survival group did this ERP effect differ in scalp distribution between high- and low-imageability words. Notably, the scalp distribution was more frontal for high-imageability words in the survival group, compared to both low-imageability words in the same group and to the analogous ERP effects in the moving group. There were no differences between groups on ROC or ERP measures of familiarity. These findings suggest that survival processing during encoding affects subsequent recollective processing and that the mechanisms underlying the effect are specific to high-imageability content.

ABSTRACT. Recognizing individual faces is a challenging task for the visual system. Recently, fMRI decoding studies have revealed face identity representations in the occipitotemporal cortex. To uncover the temporal emergence of these representations, we used representational similarity analyses (RSA) of EEG responses for familiar faces. The participants (n=26) viewed a set of ambient face images of four highly familiar celebrities (two male, two female), while performing an orthogonal task. While univariate analyses revealed a differentiation between male and female faces, they were not able to show a distinction between identities of the same sex, multivariate RSA, on the other hand, demonstrated a gradual emergence of face identity representations, with an increasing degree of invariance. We observed the rapid emergence of identity information, starting shortly after 100ms, modulated by sex differences and image similarities. From 400ms after onset and predominantly in the right hemisphere, we observed a high degree of discriminability between both opposite and within-sex stimuli, and a tolerance for image-dependent variations. As the appearance of a familiar person can vary drastically, these invariant representations may be a crucial prerequisite for successful face recognition in everyday situations.

ABSTRACT. In everyday life we usually recognise personally familiar faces efficiently and without apparent effort. This study examined to which extent the neural processes involved in recognising personally familiar faces depend on attentional resources by analysing event-related brain potentials. In two experiments, participants were presented with multiple ambient images of highly personally familiar and unfamiliar faces and pictures of butterflies, with a letter string superimposed on each image. Their task was either to indicate when a butterfly occurred (effectively ignoring the letter strings) or to indicate whether each letter string contained the letter X or N. Attentional resource load was manipulated in the letter task by presenting the target among different distractor letters (high load; Experiment 1) or by using only a single repeated letter in each string (low load; Experiment 2). ERPs revealed more negative amplitudes for familiar relative to unfamiliar faces under both high and low load conditions, both in the N250, reflecting the activation of perceptual face representations, and in the subsequent Sustained Familiarity Effect (SFE). Nonetheless, while the magnitude of the N250 effect was not substantially affected by attentional load, the SFE was still present but reduced in the high relative to the low load experiment. These findings suggest that perceptual face representations are activated independent of the demands of a competing task. However, the subsequent SFE, presumably reflecting more sustained activation needed to access identity-specific knowledge that can guide potential interactions, strongly relies on the availability of attentional resources.

ABSTRACT. Objectives: There is a considerable amount of research investigating cyberbullying although there are only a few studies using neuroimaging techniques so far. Our aim was to develop a task that is ecologically more valid than the existing social exclusion tasks (e.g. Cyberball) and allows us to measure reactions to cyberbullying with fMRI. Methods: In a pilot study, 5 university students (2 males, mean age=20.2, SD=0.84) participated. Task-related activations and deactivations were assessed using block design fMRI in the Cyberbullying Task. During the Cyberbullying Task participants see neutral posts (others’ and themselves’) and comments. There are three different conditions: (1) others get negative comments, (2) the participant gets negative comments and (3) the participant gets positive comments. In a block there are a post, six comments, and after every block there are two questions: how the one feels who gets the comments and how would the participant react to the situation. Each condition recurs five times, in all, there are fifteen blocks. Results: Right orbitofrontal cortex revealed greater and left insula showed smaller activation when participants got negative comments compared to the conditions when they got positive comments or others got negative comments. Additionally bilateral posterior cinguli showed greater activation when they got negative comments compared to when they got positive comments. Conclusions: Our results are important because insula and cingular cortex usually show activation during social exclusion/inclusion, in addition orbitofrontal cortex has an important role in self-monitoring and emotional processing. Thus our Cyberbullying Task is worthy for further research.

ABSTRACT. It has long been posited that impaired inhibitory control, similarly to other addictions, could play a pivotal role in the development and maintenance of internet addiction. Yet, to our knowledge, no studies have assessed specifically the underlying functional neural networks. Therefore, the aim of this study was to explore the relationship between internet addiction and functional networks involved in inhibitory control. Sixty healthy, young, right-handed healthy university students were included. They underwent a block-designed Stroop- and Simon-task during BOLD-contrast imaging in 3T Siemens MRI scanner. Task-related activations were assessed using model-free (T-PICA) analysis. The Problematic Internet Use Questionnaire (PIUQ) was used to measure internet addiction, and the Barratt Impulsiveness Scale (BIS-11) to measure impaired control. After visual checking of the data, we found three meaningful components that positively correlated with both impaired control subscale of PIUQ and motor impulsivity subscale of BIS-11. The activation pattern across the three components was highly similar, involving frontal areas (e.g. Precentral Gyrus, Superior Frontal Gyrus, Supplementary Motor Cortex, Frontal Orbital Cortex, etc.) as expected. Deactivations in all three components involved the Default Mode Network (i.e. Praecuneus Cortex, Postcentral Gyrus, Medial Frontal Gyrus, Frontal Pole, etc.). Our results provide further evidence for altered brain functions in internet addiction. The results may help understand the underlying factors of IA and could lend further support in developing a treatment.

ABSTRACT. Arachnophobia is a high prevalent psychiatric disorder, which is probably maintained by maladaptive associative learning. However, the neural basis of disorder-relevant associative learning in arachnophobia has not been investigated, yet. Thus, the present fMRI study explored behavioral and neural correlates of associative learning in patients (n=34) with arachnophobia compared to healthy controls (n=34). Participants performed a paradigm, in which one stimulus became disorder-relevant due to coupling with a spider picture, while another was never followed by a spider picture. Behavioral and neural responses to stimuli, which lay in the perceptual space between both stimuli, were assessed and compared between both groups. In subjective ratings, patients described the disorder-relevant stimulus and those, which were perceptually similar as more fear-evoking, arousing and negative than healthy controls. The fMRI results showed increased activity in the bilateral insula, right bed nucleus of the stria terminalis, right amygdala and left anterior cingulate cortex in patients compared to healthy controls, to the disorder-relevant stimulus and the perceptually similar stimuli compared to the stimulus, which was never followed by a spider picture. In this study maladaptive associative learning in arachnophobia was demonstrated, reflected by increased fear to perceptually similar stimuli to a disorder-relevant stimulus compared to healthy controls. Furthermore, this was accompanied by hyperactivity of the fear network.

ABSTRACT. The term ‘blindsight’ describes residual visual capacities without acknowledged awareness in the blind area of participants with visual field defects. However, some researchers have argued that blindsight might merely be only an experimental artefact caused by light scatter, i.e. light from stimuli presented in the blind field falling on areas of the retina corresponding to intact areas of the visual field. The aim of our study was to test this hypothesis in healthy participants using the blind spot to simulate a scotoma. We tested 21 participants in three paradigms: (1) movement direction discrimination, (2) redundant target task, (3) detecting a circular stimulus (temporal 2AFC). In the default condition, black stimuli were presented on a grey background. Moreover, we ran the detection task (3) in two additional conditions: white stimuli on a grey background and white stimuli on a black background in darkness. Our results show that (1) movement stimuli presented in the blind spot can be discriminated above chance at a control position but not in the blind spot. (2) In the redundant target task, a second stimulus in the blind spot did not reduce reaction times. (3) Analyzing all three light conditions of the detection task, only a white blind spot stimulus in darkness resulted in above-chance performance. In summary, light scatter can be an issue if stimuli are too bright. However, black stimuli on a grey background did not create a sufficient difference in light scatter to influence task performance and can therefore be used to investigate blindsight.

ABSTRACT. Presenting the two eyes with different stimuli results in a phenomenon called binocular rivalry (BR) as the two stimuli rival for perceptual dominance. In healthy participants salient stimuli are often dominant for longer periods of time. We tested participants with congenital prosopagnosia (cP) in a BR paradigm to evaluate whether their differences in face processing were also reflected in the BR dominance times of face stimuli. Studies suggest that the holistic perception of faces is disrupted in cP, which leads to a focus on local features of the face. This is illustrated by a finding concerning the Thatcher illusion: In this illusion mouth and eyes of a face are rotated 180°. Although the upright face looks grotesque, healthy participants often fail to notice these massive changes when the face is inverted. Participants with cP, however, also notice the changes in the inverted face. In experiment 1 we used BR to compare predominance of faces displaying different emotions (fearful, happy, neutral) vs. houses between participants with (N=21) and without cP (N=21). Results show decreased face predominance in the cP-group compared to the control group which implies reduced saliency of faces in cP. In experiment 2 we used “thatcherized” faces (vs. houses) and found a lower sensitivity to face inversion in the cP-group indicating stronger local face processing, which – as opposed to holistic processing – is more immune to inversion.

ABSTRACT. An illusory correlation (IC) is the erroneous perception that two categories are correlated. Even though numerous studies have investigated ICs, there is still debate over the exact mechanisms leading to ICs. According to the distinctiveness approach, most attention is paid to the least frequent category combination at learning. The heightened attention leads to a higher availability of this category combination and, consequently, to an overestimation of its frequency. In contrast, proponents of the accentuation hypothesis claim that attention should lie on both the most frequent and the least frequent category combination in order to maximize the differentiation between the categories. In this ERP study (N=24) we compared the distinctiveness approach with the accentuation hypothesis. An active oddball task was used to elicit a P300, a marker for subjective probability and attention allocation. If the distinctiveness account were true, there should be a linear increase of the P300 amplitude from the most frequent to the least frequent category combination. However, if the accentuation hypothesis were true, we expected P300 amplitudes to be larger for the most frequent and the least frequent category combinations than for the moderate frequent category combinations. Consistent with the distinctiveness approach, we found a linear increase in the P300 as a function of infrequency. Furthermore, a frontal slow wave differentiated between the least frequent category combination and all other category combinations. Together, our results not only support the distinctiveness hypothesis, but also indicate that a fronto-parietal network is involved in the formation of mental contingency representations.

ABSTRACT. Internet addiction (IA) is defined as an excessive and prolonged Internet using pattern that causes behavioral and cognitive problems and is generally conceptualized as a behavioral addiction. Functional brain alterations that are related to IA have been revealed, significant differences were found in brain regions involved in cognitive control and reward processing. However, other brain networks might be also altered, one possible candidate is the language system of the brain. To our knowledge, no one has investigated language networks in addictions with neuroimaging techniques, however, indirect evidences suggest its involvement. In this study, we investigated IA-related language network alterations in young adults using fMRI during phonemic fluency task in 59 healthy young adults. IA was assessed using the Problematic Internet Use Questionnaire. This questionnaire consists of 18 items, and three subscales: Obsession, Neglect, and Control disorder. To get clear picture about the phenomenon model-based (GLM) and model-free (ICA) analyses were performed. We found two independent components (IC3 & IC4) that are positively associated with control subscale of PIUQ and GLM analysis also supported our hypothesis, that language processing is impaired in IA. These results suggest that IA might have long term negative effect on neuro-cognitive processes.

ABSTRACT. Successful cognitive control during behavioral conflicts relies on neural adjustments that can occur reactively (i.e., after conflict occurrence) and proactively (i.e., in preparation prior to conflicts). While midfrontal delta/theta oscillations are known to play a role for reactive control, their relevance for proactive control is unclear. The current study tested the role of midfrontal control oscillations during the anticipation of action-related conflicts via EEG. On each trial, participants (N = 33) either had to perform a frequent, prepotent action or an infrequent, conflict-inducing alternative response. This action was preceded by a cue, which either reliably indicated the absence of any action conflict (certain-cues) or that a conflict might potentially occur (maybe-cues). We investigated both inhibition conflicts (suppression of prepotent actions) and switching conflicts (alteration of prepotent actions). Results showed clear evidence for proactive adjustments in neural activation prior to the action onset: Maybe- compared to certain-cues led to significantly stronger suppression of posterior alpha oscillations during preparation, indicating intensified sensory processing. Additionally, maybe- compared to certain-cues led to task-dependent preparatory adjustments of motor activation, as indicated by increased central beta suppression during switching conflicts, but decreased beta suppression during inhibition conflicts. Importantly, while midfrontal control oscillations significantly increased after a conflict occurred, preparatory neural adjustments during conflict anticipation were not accompanied by increases in proactive midfrontal oscillations. This dissociation suggests that midfrontal oscillations are an important part of reactive adjustments during behavioral conflicts but that proactive adjustments of neural processes can be implemented without midfrontal oscillatory control.

ABSTRACT. There has been a long debate on whether emotional stimuli like angry faces or emotional scenes draw attention and gaze faster than neutral stimuli. Previous studies using physically salient stimuli showed that distractors lead to slower eye-movements and slower neural responses towards targets (Kulke, Atkinson & Braddick, 2015). The current study aimed to investigate the effect of emotional salience by combining eye-tracking and electroencephalography (EEG) to measure neural mechanisms of overt attention to faces displaying emotional (happy or angry) or neutral expressions. Thirty-four participants completed a gaze-contingent fixation shift paradigm (preregistration: osf.io/vbk2e). Once they fixated on a central dot, faces appeared in the periphery and disappeared once participants moved their eyes towards on them. There was no effect of emotional expression on eye-movement latency or P1 latency in response to the peripheral face, which was in contrast to previous studies manipulating physical rather than emotional salience. However, the amplitude of the later EPN response was significantly larger to emotional compared to neutral faces. This component occurred after the saccade towards the stimulus had been completed. In summary, modulations of event-related potentials by emotional content only occurred after an eye-movement towards the stimulus. Therefore, emotional peripheral stimuli did not draw attention faster than neutral ones. Fast eye-movements may be guided by physical rather than emotional salience in the fixation shift paradigm.

ABSTRACT. An influential model of vision posits two streams of visual information – one dorsal and crucial for localising visual objects (e.g. for actions, such as eye movements), the other ventral and crucial for identifying visual objects. In line with this dichotomy, ventral processing of visual objects has often been characterized as location-invariant. However, recent studies found that typical face-directed gaze behaviour reflects location-dependent biases in perceptual sensitivity (de Haas et al., 2016 JNeuro; de Haas & Schwarzkopf, 2018 JoV) and may be linked to corresponding biases in neural tuning within the inferior occipital gyrus (IOG).

Here, we present data from an fMRI experiment directly testing the hypothesis that neural tuning for facial features is linked to spatial (i.e. retinotopic) tuning in IOG (n = 14 hemispheres). We used encoding models to probe voxel-wise spatial preferences and independently tested the preferred relative position within a face. The majority of responses were well explained by Gaussian population tuning curves (pTCs) for spatial location and face-parts. Parameter maps revealed a common gradient of spatial and face-part selectivity, decreasing from posterior to anterior IOG. Preferred location was organised more idiosyncratically, but showed local clustering and was correlated across maps of visual and face space.

These findings reveal correlated spatial and face-part maps as principles of functional organisation in IOG and thus a tighter link between ‘what’ and ‘where’ than proposed by the dual-stream hypothesis. This organisation likely reflects developmental boundary conditions, constrains the neural mechanisms of face perception and typical gaze behaviour.

ABSTRACT. The grounding of concepts in the sensorimotor brain systems is discussed controversially. Modality-specific models propose that concepts are embodied in the sense that they are represented in distinct sensory and motor brain areas depending on specific sensory and motor experiences during concept acquisition. Accumulating evidence suggests that concrete concepts are closely linked to the sensorimotor systems, whereas the mere existence of abstract words seems to contradict embodied approaches. Here, we adopted a theory-driven approach frequently used in the investigation of concrete concepts to the domain of abstract concepts and compared brain activation to well-defined subtypes of abstract concepts with a known feature content using functional magnetic resonance imaging. Carefully matched visual (e.g. beauty) and motor (e.g. fitness) abstract words were presented to n = 24 participants along with pseudowords while performing a lexical decision task. Furthermore, participants performed two localizer tasks by actually moving their hands and by looking at real pictures. Processing of visual abstract words specifically elicited higher activity in occipital visual areas whereas processing of motor abstract words specifically activated frontal and parietal motor areas. In addition, these differential fMRI signals showed overlapping patterns with brain activations observed during the motor (pre- and postcentral gyrus) and visual (fusiform and lingual gyrus) localizer tasks. Consistent with the grounded cognition framework, our results suggest that, similar to concrete concepts, visual and motor abstract words are grounded in the corresponding modal brain systems typically engaged in actual perception and action.

ABSTRACT. Previous research has shown that the cardiac defense response (CDR) is enhanced when participants perform simultaneously an external attentional task. The aim of this study was to examine autonomic mediation in this type of modulation. Participants were 30 university students (20 women and 10 men) randomly assigned to two conditions. The psychophysiological test consisted of two presentations of an acoustic stimulus capable of eliciting the CDR with an inter-trial interval of 12.5 min. Participants belonging to Condition 1 performed Sternberg's Visual Search Task using neutral pictures as targets. The task had 80 s duration and started immediately after each noise presentation. Participants belonging to Condition 2 performed no task. Continuous measurements of heart rate (HR), pre-ejection period (PEP), and systolic blood pressure (SBP) were used. The dependent variables were the simultaneous beat-to-beat pattern of HR, PEP, and SBP. Results concerning the CDR showed a reduction of the first decelerative component and a potentiation of the second accelerative component in the attentional group, compared to the control group. Regarding autonomic mediation of the CDR, it was observed a higher PEP and a lower SBP for the attentional group, coinciding temporarily with the second accelerative component of the CDR. These results suggest the involvement of the parasympathetic branch in external attentional modulation of the CDR.

Funding: This work was supported by the MINECO/AEI/FEDER, UE [project number PSI2017-83007-P] and the MECD [grant numbers FPU15/03621, EST17/00790].

ABSTRACT. The connection between the cognitive functions like verbal performance and the anxiety have been shown in behavioral studies. The level of anxiety correlated negatively with the result of verbal (letter and word) fluency task. The aim of our study was to measure the neural background of the connection of the verbal performance and personal differences in anxiety. We hypothesized that the activation of Broca-related language networks are affected by the level of trait anxiety. Ninety healthy university students (45 males) were included in our study, to investigated anxiety-related language network alterations during a verbal fluency task. The self-reported anxiety questionnaire was used to measure trait anxiety. Task-related activations and deactivations were assessed using block design fMRI in a phonetic verbal fluency task, with seven cycles of 30 second long altering rest and word generation task. During the verbal task the subjects were asked to silently generate different words with a particular letter without any movements. During the rest periods, they were instructed to stop the active task and relax. The measurements were performed on a 3T Magnetom Tim TRIO, human whole-body MRI scanner (Siemens AG, Erlangen, Germany) with a 12-channel head coil. Model-free independent component analysis (ICA) was used on fMRI data to assess task-related networks. We found 21 task related meaningful components 6 of which correlated significantly with trait anxiety. The activation measured in the Broca area confirm our hypotheses. According to our results the language networks are less pronounced in people with higher levels of anxiety.

ABSTRACT. The question whether spatial selective attention is necessary in order to process vocal prosody has been controversially discussed when testing sighted individuals. Whereas some studies argue that spatial attention is necessary in order to process emotional human voices, other studies point to the fact that emotional human voices (e.g. anger) can be processed even outside the focus of spatial attention (e.g. Grandjean et al., 2005). Here, we asked if longterm visual deprivation starting from birth requires attention in order to process emotional information in human voices.Therefore, eight congenitally blind individuals and thirteen sighted controls had to attend either to the left or to the right loudspeaker and to detect rare deviant syllables spoken in neutral, angry, happy, and fearful prosody while the EEG was recorded. Blind individuals processed emotional voices at the attended speaker more efficiently compared to sighted controls. In contrast to sighted controls, who showed early attention effects in the time range of the N1 in the fearful prosody but not in the neutral, happy, and threatening vocal prosody (manifested in an interaction between emotion and attention), blind individuals do not show an enhanced capture effect in the fearful condition compared to the other emotions (main effect of emotion, main effect of attention). Moreover, blind individuals show enhanced N1 amplitudes irrespectively of the presented emotion or the attention conditions. Enhanced excitability of unimodal sensory areas might be one underlying neural mechanism responsible for emotional voice processing in congenitally blind individuals.

ABSTRACT. The testing effect refers to the finding of facilitated learning in conditions in which participants study and receive practice tests with conditions in which they receive an equivalent amount of further studying in lieu of practice tests. In more educational terms the testing effect can be compared formative assessment which is used to monitor student learning to provide ongoing feedback. In the present study we combined the testing effect with the provision of timely feedback during an artificial language learning task. 24 participants studied 72 grammatically sentences of the artificial language BROCANTO to learn the underlying grammatical rules. One group of participants received additional 72 sentences for further study and a second group received a different set of 72 sentences, one-half of which were ungrammatical, making grammaticality judgments on each sentence presented. The third group received the same test as the second group but received additional feedback on each judgment. Results replicated the typical testing effect demonstrating an advantage of restudying over testing on a final test immediately after the initial learning but the opposite effect at a retention interval of 48h. Crucially, feedback enhanced performance at both retention intervals with larger improvements at 48h. These results indicate that the testing effect is underlying the benefits of formative assessment that can be further improved by feedback.

ABSTRACT. Reading digital texts is a common practice in today’s education. Prior studies showed that the coherence of a text can influence text comprehensibility with low degrees of coherence causing attention failures (mind wandering) and, consequently, negatively impacts reading comprehension. In addition, working memory capacity (WMC) and prior knowledge of the subject have been suggested to be related to both reading comprehension and mind wandering. However, results remain controversial as the interaction of these three factors has not yet been explored. 85 participants either studying law or a different subject read either a coherent or incoherent version of the same unfamiliar hypertext about the copyright law. While reading, they reported self-caught mind wandering with task-embedded thought probes. After reading the hypertext, subjects were probed on their text comprehension. Supporting prior findings, mind wandering did occur more frequently when participants read difficult rather than easy texts regardless of their undergraduate course. Moreover, this was modulated by WMC in that participants with lower WMC exhibited more frequent mind wandering than high WMC participants solely when reading low coherent texts. In addition, high WMC participants outperformed low WMC participants on all measures of text comprehension. With a low WMC it seems difficult to inhibit irrelevant information and access related information from working memory, especially when text complexity is high. Interestingly, the present results also indicate that prior knowledge benefits later text comprehension despite not affecting reader’s attention. These findings provide insights into processing attention during reading online texts

ABSTRACT. When naming a sequence of pictures of the same semantic category (e.g., clothes), latencies increase with each additional picture. According to one account, this cumulative semantic effect is assumed to reflect interference during lexical access (e.g., Howard et al., 2006), and seems to be modulated by semantic similarity (Rose & Abdel Rahman, 2017). While this effect has been shown for morphologically simple nouns (e.g., basic level category members, such as blouse or pants), it is yet unclear if and how the semantic interference differs for semantically transparent compound nouns, which are subordinate category members (e.g. silk blouse or leather pants). In the current (ongoing) study, we use the continuous picture naming paradigm to investigate the difference between these two types of nouns. Participants (N=36) are presented a series of objects that are derived from different semantic categories, and are asked to either use simple nouns or compounds when naming these out loud. Although we expect to find a cumulative semantic effect in both conditions, we predict weaker interference for the compounds. Our simple-noun targets are basic-level category members that share most common semantic attributes (Rosch & Mervis, 1975), whereas the compounds are subordinate category members that are semantically richer but less interconnected. Therefore, we predict stronger co-activation and, thus, stronger interference for the simple-noun than the compound targets.

ABSTRACT. The literature on the embodied simulation view suggests that language comprehension activates traces of related motor experience. The reading-by-rotating paradigm introduced by Zwaan and Taylor (2006) for instance uncovered a compatibility effect between the content of a sentence and the actual movement participants performed during reading. Participants read sentences by rotating a knob either clockwise or counterclockwise (e.g. Eric turned down/up the volume.). Reading times were faster when the described actions involved a direction of rotation matching the required direction of rotation of the knob. A similar study by Claus (2015) with verb-gapping sentences (e.g., John closes a juice bottle and Jim a lemonade bottle.) produced the opposite pattern of results, with an advantage for mismatching rotation directions. Claus (2015) discussed several potential accounts for the deviances in the results, referring either to differences in the devices used for rotation, or to differences in the materials (one vs. two characters). The current replication study aims at further investigating these accounts. We used the exact same device as in Claus (2015), but our sentences always described only one character. A preliminary analysis of the results (N=43 out of 80 participants) shows numerically faster response times in the mismatching condition, similar to the result obtained by Claus (2015). In the current analysis, this difference is not significant. However, if it turns significant in the final analysis, this would undermine the idea that the number of characters in the sentence is the relevant factor for obtaining one or the other result pattern.

ABSTRACT. In experiments investigating the processing of true and false negative sentences, it is often reported that polarity interacts with truth value, in the sense that true sentences lead to faster reaction times than false sentences only in affirmative conditions. Also in electrophysiological data, negative sentence typically do not show the expected facilitation in true versus false conditions. Various reasons for this difference between affirmative and negative sentences have been discussed in the literature (lexical associations, predictability, informativeness).

In the present study we excluded lexical associations as a potential influencing factor. Participants saw artificial visual worlds (e.g., a white square and a black circle) and the corresponding sentences (i.e., “The square/circle is (not) white”). The results again showed a clear effect of truth value for affirmative sentences (true faster than false) but not for negative sentences. This result implies that the well-known polarity-by-truth-value interaction cannot solely be due to long-term lexical associations. Additional predictability manipulations allowed us to also rule out an explanatory account that attributes the missing truth value effect for negative sentences to low predictability. In our presentation, we will discuss the viability of an informativeness account.

ABSTRACT. Absolute adjectives, in contrast with relative adjectives (tall/short), are said to behave symmetrically, the negation of one member entailing its pair: not dead=alive and not alive =dead. Recent studies have shown that absolute adjectives are rather asymmetric: right ≠ not wrong and bound ≠ not (Kennedy and McNally 2005; Paradis and Willners 2006). This talk aims at discussing the negated absolute adjectives in Romanian. To that end, a judgment test was designed in which the sentential negation nu A ‘not A’ (She is not healthy) was tested against the affixal negation neA ‘unA’ (unhealthy), the affixal negation non-A (non-healthy) and the polar opposite B (sick). The negative affix ne- “un-” is said to express a scalar, ‘neither-nor’ reading (which is incompatible with absolute adjectives). We hypothesize that if the participants choose this negative affix as the interpretation for the nu A ‘not A’, then the adjectives have scalar features which lead to the asymmetries. 18 absolute adjectives forming 9 canonical pairs were tested (N=66). The Principal Component Analysis was used to analyze the data (Rstudio software and FactoMineR and Factoshiny packages). The results show different profiles of the adjectives, displayed in three categories: (a) absolute adjectives which have different degrees of scaling potential, (b) relative scalar adjectives, with some absolute traces and (c) adjectives that can be both absolute and relative, their interpretation being set contextually. This cast new light on the interpretation of canonical antonym pairs as well as on the absolute interpretation of these adjectives.

ABSTRACT. Research with bilinguals revealed certain differences in the way a native language (L1) and a language that is learned later as a foreign language (L2) shape our thoughts. For example, moral evaluations (Hayakawa et al., 2016) or risky decisions (Costa et al., 2014) were found to be more rational in L2 than in L1. Our research question is centered around the processing of swear words presented either in L1 or L2. Bowers and Pleydell-Pearce (2011) recently showed that pronouncing swear words (e.g. Fuck) in one’s native language provokes stronger electro-dermal activity (EDA) than pronouncing its euphemisms (e.g. “F-word”). This, although both verbal forms refer to the same concept. The authors thus suggested that some form-affect associations might be established bypassing the semantic system. In the present study we borrowed Bowers and Pleydell-Pearce’s paradigm to contrast L1 and L2. Native German and Turkish participants named swear words and neutral control words in their native language as well as in their foreign language. EDA and pupil sizes served as dependent measures. Our findings are discussed within the frame of linguistic relativity.

ABSTRACT. Bi- and Multilingualism are concepts that do not hold a special scarcity value in today’s world anymore. With continuous globalization and immigration even mainly language-homogenous countries show increasing numbers of children and adults whose dominant language differs from the language of the country. In this scenario the question arises how education has to be structured to be most efficient in bilinguals with a distinct language dominance and if bilingual education might be beneficial? In the following study we aimed to get more insight into the possible importance bilingual education can have in mathematics education. Our research question was whether bilingual education is helpful to bilinguals when trying to learn arithmetic facts in their less dominant language. We tested 90 bilingual adults whose first and dominant language was German and who spoke English as their second language. Participants were assigned to three different training groups and either practiced simple multiplications in German, English or both languages on five consecutive weekdays. Reaction times on solving multiplications in both German and English were recorded before and after the training. The results showed that language dominance seems to have a major effect in educational situations. German items were solved more quickly than English counterparts, before and after the training and in every training group. Nevertheless, there were more learning gains for English items when participants received a training in both languages. This lets us conclude that bilingual education can be beneficial when bilinguals are learning arithmetic facts in their less proficient language.

ABSTRACT. This poster focuses on the difficult domain of Aktionsart and examines whether some of its dimensions could be grounded to captured motion data. This analysis and the visualized plots that were exported from various actions offer a sound basis of aspect’s understanding in language and motion. Aktionsart is a multidimensional linguistic phenomenon, which encodes temporal and frequency information. It is considered to play significant role to mental simulation of an action both in the execution of the movement -per se- and the linguistic expression of the real world actions. According to Talmy (2000, II), languages form event temporal information into a particular typology according to the manner of the change or not state and especially the pattern of distribution along time. The graphic representations of the lexical aspect have many similarities with those of the kinetic data. Preliminary results are: • Elicitation of features (duration, iteration) resembling Aktionsart and plots of path, direction, velocity (Scheme 1). • Correlation between kinematic features and linguistic characteristics Apart from the kinematic features that are crucial for human motion analysis, anaphora resolution (gestures), machine translation (for ASL) etc, we wanted to directly link sensorimotor and linguistic information. Running correlation matrices for kinematics and linguistic variables-taken from corpus driven analysis-we observed that the two significant correspondences are (i) between variables verb_class (it groups motion verbs according to Levin’s (1993) classes) and repetition and (ii) repetition with linguistic duration, a notion that implies Aktionsart but it does not carry the same semantic load.

ABSTRACT. According to the experiential-simulations view of language comprehension (Zwaan & Madden, 2005), we comprehend the meaning of words by re-activating sensorimotor experiences that are associated with the word’s referent. The work of Lachmair, Dudschig, De Filippis, de la Vega, and Kaup (2011, Experiment 2) provides positive evidence for this claim with respect to spatial experiences, indicating that the associated location of a referent in vertical space is automatically re-activated upon hearing or reading the corresponding word. Their participants responded to words such as “bird” or “worm” (that are associated with an upper or lower vertical location, respectively) by performing an upward or downward hand movement depending on the font color of the words. Responses were faster when the hand movement matched the referent’s associated vertical position. In our current work, we aimed at developing an online paradigm for replicating these findings. Instead of a vertical response device we used standard keyboard and mouse responses. In Study 1 (N = 56), participants clicked on the word that was presented in the middle of the screen and dragged it to the upper or lower part of the screen by using their mouse. In Study 2 (N = 50), keyboard response keys were associated with an upper or lower location by means of implicit learning. Data analysis showed that responses in Study 1 as well as in Study 2 were significantly faster in compatible than in incompatible trials. We conclude that an actual motor movement is not essential for observing the compatibility effect.

ABSTRACT. The complex interplay of prosody and semantics and their role in the perception of emotion in spoken language is a topic open to discussion. Different approaches try to determine the nature of this interaction, for instance which factor dominates over the other. The present study provides the means to examine precisely this interaction in German by generating sentences with validated emotional content regarding the emotions anger, fear, happiness, sadness, or neutral (expressing no emotion) in both semantics and prosody. German native speakers (N = 61) selected the sentences out of a 400 sentences comprising collection based on ratings on a 6-point Likert scale in terms of emotional semantic content. The result thereof are 55 sentences, 11 sentences in each case could reliably be associated with one of four distinct emotions or the neutral category. The sentences were recorded in different prosodies, either congruent to the semantic content or incongruent, by a professional actress. A group of experts such as speech therapists and researchers rated these recordings on the clarity of speech and emotion conveyed via prosody. The result is a compilation of 50 German sentences with separately validated emotional content in prosody and semantics. Moreover, the sentences are carefully balanced regarding linguistic factors (word frequency, phonological neighborhood density, and number of syllables). The linguistic balance enables an unbiased evaluation of the roles of semantic content and prosody in emotional speech. This allows an independent variation of prosody and semantics in studies with factorial design.

ABSTRACT. Evidence for a mechanism that validates incoming text information based on prior knowledge and beliefs accumulates, yet the role that source information might play for validation is unclear. We investigated how text-belief consistency and source information could affect validation of short authentic texts presented as Twitter messages (tweets). Sixty-four participants read 64 tweets about four socio-scientific, conflicting topics (e.g. climate change) stated by different sources in a self-paced fashion. Each text represented one of two argumentative positions regarding the topic. The sources were consistent or inconsistent with the argumentative position of the message (source-message consistency), resulting in two versions of each item. Participants’ pre-measured beliefs were used to determine the consistency of each message with individual beliefs (text-belief consistency). Participants read all tweets (plus 64 fillers) in a randomized order and answered comprehension questions after filler tweets. Reading times and (in a separate block) ratings of plausibility and source credibility of the tweets were collected. The results revealed text-belief consistency effects for text-belief and source consistency. Participants read belief-consistent messages faster than belief-inconsistent messages and messages from sources consistent with the argumentative position faster than messages from inconsistent sources. Moreover, for the ratings of plausibility and source credibility, interaction effects of text-belief consistency and source consistency emerged: Source-message consistent and text-belief consistent messages were rated as more plausible and its sources as more credible than inconsistent combinations. We conclude that not only text-belief but also source information are taken into account while validating multiple short texts about complex topics.

ABSTRACT. Dialogue systems for fault diagnosis can query users to elicit problem descriptions incrementally, but speech is inherently ambiguous. This study investigates whether explanations of question intent can prevent misunderstandings. Eighteen participants were instructed about the production and packaging of chocolate, enabling them to understand fault causes. During the experiment, they were asked whether a given description matched the picture of a faulty chocolate bar or not. In three blocks of a within-subjects design, participants either received only questions, questions accompanied by explanations of question intent (i.e., which cause the system wanted to infer from their response), or questions, explanations, and the option to clarify their response and thereby prevent incorrect interpretations by the system. After each trial, participants rated their certainty of having understood and answered the question correctly. Correctness, certainty ratings, solution times, and the content of participants’ clarification texts were analyzed. With explanations, participants responded more correctly than without. However, this explanation benefit was absent when participants had the opportunity to clarify, suggesting that they were less likely to adopt the system’s wording. Roughly half of their texts referred to fault causes, while the other half provided reformulations. Confidence ratings were unaffected by explanations, and solution times were increased both by the presence of explanations and the opportunity to clarify. Taken together, the results imply that explanations support the prevention of misunderstandings. However, when given the chance, users accomplish this by providing additional clarifications rather than adopting the words used by the system.

ABSTRACT. Compound words can be semantically transparent when the contribution of their constituent meanings is clear (airport), but the degree of semantic transparency varies between compounds (compare strawberry, staircase, ladybird). Semantic transparency has been shown to affect compound processing and comprehension in a wide range of tasks. Critically, recent studies have demonstrated that semantic transparency has to be operationalized as a multi-dimensional construct, comprising semantic relatedness (how similar is the meaning of straw to strawberry) as well as compositionality (how well is the meaning of strawberry predicted from straw and berry, and what is the contribution of straw to the predicted meaning?). Here, we present a comprehensive and multi-dimensional dataset of semantic transparency measures for 1,810 German compounds, obtained from a computational and fully implemented semantic model. These measures are validated using data from four behavioural experiments: Explicit transparency ratings (n = 26), two different lexical decision tasks using different nonwords (both n = 16), and an eye-tracking study (n = 8). We demonstrate that different semantic effects emerge in different behavioural tasks, which can only be captured using a multi-dimensional approach to semantic transparency.

ABSTRACT. How do people process and evaluate falsehood of statements? The congruency-hypothesis assumes that processing false sentences is not different from processing valid sentences, and accordingly, activating any content-related information – even if it is false in the context of a sentence– should facilitate processing and validation. Alternatively, the validity-hypothesis assumes that validating false sentences is unique, and necessitates activating content that makes false sentences valid. Participants were asked to validate sentences which were preceded by images that conveyed the component of the statement that either made it valid or false (e.g., an image of tracks/highway preceding the sentence “trains run on tracks/highways”). Results from three experiments support the congruency-hypothesis, demonstrating that activating congruent concepts facilitates validation for both false and valid statements. The findings accordingly challenge the assumption that activation of valid content is necessary for detecting falsehood. Rather, a detection of falsehood seems to rely on the association of a sentence’s content to particular semantic networks.

ABSTRACT. Source reinstatement facilitates source retrieval

In the classical source monitoring paradigm, source memory failures can result from lack of storage or retrieval of source details. Based on previous research (Dodson & Shimamura, 2000), we suggest that source reinstatement, that is, reinstating the original source at test, facilitates source retrieval. To test this assumption, we extended the classical paradigm by a second test that directly followed the first, standard source test. Specifically, all target items and distractors from test 1 were tested again, this time however, by both sources consecutively (e.g., first spoken by Source A, then by Source B), such that for originally studied items, one presentation was a reinstatement of the original study context. We additionally varied difficulty of encoding and difficulty of reinstatement by manipulating presentation time (i.e., repeating vs. not repeating each word) and source similarity (i.e., using a distinct female and male voice vs. two male voices) between subjects during study. Multinomial model-based analyses indicate a clear source memory enhancement in the retrieval-facilitating second test compared to the first test, however, primarily for the high-similarity groups. Crucially, repetition had also a beneficial effect on source memory, but only in test 2. Because repeating a word during encoding should primarily influence source (and item) storage, this finding corroborates our assumption that source memory in test 2 mostly depends on storage. Altogether, our results suggest that source reinstatement is a useful technique to facilitate retrieval of stored source details, especially if source discrimination is difficult.

ABSTRACT. Is Source Guessing Automatic or Controlled? Examining Cognitive Load and Aging

According to the probability-matching account, prior knowledge (e.g., stereotypes) biases source guessing when people cannot remember the source of information and accurate perception of item-source contingency is impeded. However, little is known about the nature of underlying mechanisms that qualify stereotype and contingency-based source guessing. Thus, the main research question is whether the process of source guessing is rather automatic or controlled. We examined the nature testing the effects of cognitive load in younger adults (18 – 26 years) focusing on source guessing of age-stereotypic information. We manipulated provision of specific source information (age; at encoding vs. at test) and cognitive load in the test phase (load vs. no load) between subjects. Preliminary data based on the two-high-threshold multinomial model of source monitoring of 113 participants point towards a replication of the probability-matching account: Under full attention at test, participants who knew the sources’ ages at encoding guessed the source based on the item-source contingency; participants who learned the sources’ ages only at test guessed based on age stereotypes. Testing the nature of stereotype-based source guessing comparing participants under full and divided attention at test revealed that source guessing tended to be similarly pronounced reflecting an automatic, resource-independent process. Contingency-based source guessing seemed to counteract stereotype-based source guessing automatically, as it was comparable between participants under full and divided attention when providing source ages at encoding. Thus, both stereotype and contingency-based source guessing reflect automaticity rather than controllability.

ABSTRACT. Judgments of Guessing Partially Correct the Expectancy Illusion on Judgments of Source Schematic expectations affect source monitoring. When asked to remember the source of items, people more often correctly attribute items to schematically expected than to unexpected sources (e.g., oven in the kitchen vs. hairdryer in the kitchen; expectancy effect). However, this performance advantage stems from schema-consistent source guessing (e.g., Schaper, Kuhlmann, & Bayen, 2018). In fact, pure measures of source memory often show an inconsistency effect (i.e., enhanced memory for unexpected vs. expected sources). In our past studies, people incorrectly predicted an expectancy effect on source memory when providing Judgments of Source (JOSs). In the current research, we investigated whether this apparent dissociation between memory and metamemory exists because people factor their schema-consistent source guessing into their memory prediction, resulting in (accurate) predictions of overall performance. We hypothesized that the expectancy effect on JOSs decreases if participants also provide Judgments of Source Guessing (JOG). Seventy-two participants studied expected and unexpected source-item pairs. Half of the participants provided item-wise JOSs, and the other half both item-wise JOSs and JOGs. In JOGs, participants accurately predicted a guessing advantage for expected pairs. In both groups, JOSs showed an illusory expectancy effect. However, when JOGs were additionally provided, this expectancy effect was smaller. Thus, guessing convictions influenced metamemory judgments when not assessed separately. Nonetheless, although people seemed to be somewhat aware of their schema-based source-guessing bias there remained a robust metamemory illusion regarding expectancy effects on source memory.

ABSTRACT. Adaptive memory: Source memory is better for animate than for inanimate entities

The animacy effect refers to enhanced memory for animate over inanimate items. Two experiments test whether this memory advantage generalizes to source memory. A multinomial model of source memory was used to disentangle item recognition, source memory, and guessing processes. In Experiment 1, animate and inanimate words were presented at different spatial locations on the screen. Source memory for the spatial locations of the items was enhanced for animate compared to inanimate items. In Experiment 2, pseudowords were associated with animate and inanimate properties. Replicating previous results, pseudowords were better remembered when associated with animate properties than when associated with inanimate properties. Also, participants had better source memory for the association of pseudowords with animate properties than for the association with inanimate properties. These results strengthen the idea that animate items are associated with richer mnemonic representations than inanimate items.

ABSTRACT. The influence of social relevance on source memory and the influence of trustworthiness on behavior

Previous research has demonstrated that people remember negative reputational information particularly well. A popular assumption in evolutionary psychology claims that reciprocal altruism is supported by a cognitive module that helps individuals to remember defectors. Recent findings indicate that source memory for cheaters is not generally enhanced. It could be shown that persons have better source memory for expectancy-incongruent information than for expectancy-congruent information. In two studies, we wanted to investigate the influence of relevance on source memory for defectors. Furthermore, we were interested in the question if this source memory advantage can influence behavior. In study 1, we examined the influence of the social situation on memory for social-exchange relevant information. Faces were shown together with descriptions of cheating and trustworthy behavior. In addition, the importance of the social situation was manipulated: Participants had either to decide if they would want to work with the described person on a student project (socially relevant situation) or if they would want to ask this person what time it is while waiting on an airport (socially irrelevant situation). Results showed that only in the social relevant context a source memory advantage for cheaters was found. In study 2, we were interested if this source memory advantage for defectors can influence later behavior. Results indicated that participants spend more money when their opponent showed trustworthy behavior before. However, the appearance of the opponents seems to influence these choices.

ABSTRACT. On the musically beautiful – Revealing the physical, physiological and psychological determinants of beauty judgments of music

Empirical aesthetic research has searched for the physical and psychological reasons for why humans have always sought and thrived to produce beautiful artifacts. This human search for beauty applies to sounds too: In music, evaluative processes, such as judgments of beauty or liking, are a common part of human daily behavior. In spite of this commonality, these evaluative processes for music have received less attention than similar processes related to arts. In an acoustic, behavioral and neuroimaging study, we aimed to fill this gap by investigating the universal factors driving beauty judgments with a naturalistic free-listening paradigm. In two experiments, one group of participants continuously rated the perceived beauty of three musical pieces with a motion sensor and another group had their brain activity measured with functional magnetic resonance imaging (fMRI) while they listened continuously to the same three musical pieces. The continuous behavioral ratings were used to identify the music passages that were most consistently evaluated as beautiful/ugly across participants and to then set up the fMRI analysis of the universal neural responses to musical beauty. A third experiment with another group of participants (music composition experts) aimed to identify the musicological features that characterized the consistently-judged beautiful passages. Evidence from fMRI analysis revealed focal medial orbitofrontal activity in response to the consistently-rated beautiful passages in the musical pieces.

ABSTRACT. Aesthetic experience of music: expectation, emotion, complexity and pleasure

Aesthetic experience is the psychological process of finding satisfaction in the form and content of a piece of music (Levinson, 2009). To elucidate the aspects of form and content involved, this research tests the hypothesis that aesthetic experience arises from the generation of expectations for forthcoming musical events and their subsequent confirmation or denial which lead to experiences of tension and resolution (Meyer, 1956). These expectations are held to arise from a process of probabilistic prediction based on an internal cognitive model of the syntactic regularities of a musical style that is acquired through a long-term process of statistical learning. Empirical studies have shown that listeners’ continuously varying emotional state of arousal and valence can be predicted by an information-theoretic computational model of these processes of statistical learning and probabilistic prediction (Egermann et al., 2013; Gingras et al., 2015; Sauvé et al., 2018). Going beyond emotional experience, two recent empirical studies demonstrate that the same model accounts very well for the perception of musical complexity, providing a basis for testing the inverted-U shaped relationship between complexity and aesthetic appreciation hypothesised by Berlyne (1974), amongst others, using a principled model-based approach with real music. The results of an experiment in which 44 participants made continuous liking judgements while listening to 55 pieces of music provide evidence of the hypothesised quadratic relationship between information-theoretic complexity and aesthetic appreciation.

ABSTRACT. The gap between aesthetic science and aesthetic experience

This talk is based on my recent paper of the same name1. Scientific aesthetics sometimes uses the psychophysical approach: We vary an objective property of the stimulus (the X-axis) and measure the subjective response (the Y-axis). However, real aesthetic experience involves rare and special emotional states, such as ‘aesthetic rapture’ or a ‘sense of the sublime’. Unfortunately, such ineffable feelings cannot be evoked or measured, and cannot be plotted on our Y-axis. More ordinary emotions, such as sensory pleasure, are also part of the aesthetic experience. But in practice it is very difficult to obtain any hot reaction with well-controlled stimuli over repeated trials. Thus we resort to plotting cold evaluations on the Y-axis. Now the X-axis becomes deeply problematic. Aesthetic experiences are triggered by gestalts, or wholes, where many dimensions work together (perhaps in harmony or intriguing contrast). Thus preferences for individual features cannot say anything about attraction to real things in art or nature. This ‘Gestalt Nightmare’ was demonstrated in an experiment with just three dimensions: symmetry, color and curvature. I suggest that the future of scientific aesthetics requires transcending the quasi-psychophysical approach, but then philosophical aesthetics also has obvious and well-known flaws, such as difficulty adjudicating between contradictory claims.

1. Makin, A. D. J. The Gap Between Aesthetic Science and Aesthetic Experience. J. conciousness Stud. 24, 184–213 (2017).

ABSTRACT. Extending previous work by our group (Knoop, Wagner, Jacobsen, & Menninghaus, 2016) on the conceptual space of aesthetic appreciation terms, we obtained new data for sixteen additional object classes (visual arts, architecture, landscapes, advertisement, websites, theater, movies, music, musicals, dance, movements, faces, touch, smells, food, and wine). We first analyzed the new data in conjunction with our previously published data on literature. Cluster analyses and a network analysis based on the German raw-data for altogether 22 domains yielded similar results: three clearly separated object classes (smells & taste related object classes, faces, landscapes) on the one hand, and a much more interconnected group of man-made object classes and artistic domains on the other. In a second step, we also integrated all comparable free listing data reported by other groups into a comprehensive analysis covering 29 domains. The 29 domains/ object classes cluster consistently in groups (narratives, human-made artefacts, faces & pattern, landscapes, sound & movement, smells), reflecting sensory modalities (vision, audition, olfaction, multi-modal) as well as the distinction of natural objects vs. artefacts. Notably, ‘beautiful’ is the only term listed for all domains/ object classes. Thus, our overall analysis confirms the central role of beauty in aesthetics.

ABSTRACT. Emotion matters: different psychophysiological responses to expressive and non-expressive dance movements

The art form dance utilises this communicative power of the human body to convey meaning through movement alone. However, since the 1950ies, formalist dance has gained momentum. Here, the most important aspect is the beauty of the lines and shapes that the dance movement draws in space – independently of any expressivity. The debate between expressive and formalist dance raises an interesting empirical question for neuroaesthetics: Two dance movement sequences being equal, but one with and one without expressivity – which sequence would lay audiences prefer? And do interindividual differences modulate these preferences? Fourty-one participants watched and rated 40 pairs of short dance videos in randomised order (N=80). Of each pair, one version of the movement sequence was emotionally expressive (clip a), while the other version of the same sequence (clip b) was not expressive. Participants rated expressivity (part 1), and how much they liked each movement (part 2), while their galvanic skin response (GSR) was recorded online. Interindividual difference measures included interoceptive accuracy and heart rate variability. Participants found expressive dance clips more expressive and liked expressive clips more than the non-expressive clips. Besides, participants’ GSR differed, depending on the category of clips they were watching and this relationship was modulated by interceptive accuracy and heart rate variability. Overall, these results provide first evidence that the intended expressivity of the dancer in a dance movement influences the aesthetic experience that audiences may derive from a dance movement.

ABSTRACT. Rhythms in the signal and rhythms in the head

The brain has rhythms - and so do music and speech. Recent research has revealed that the temporal structure of speech and music and the temporal organization of various brain structures align in systematic ways. The role that brain rhythms play in perception and cognition continues to be elucidated through studies of various types. I will describe some results that illuminate the temporal structure of perceptual experience and how this may form the basis for aesthetic experience. From recognizing speech and melodies to building abstract mental structures, how the brain constructs and represents time reveals unexpected puzzles.

ABSTRACT. Let Me Read You a Story: Effects of Vocal Delivery of Literary Texts and Mood on Aesthetic Emotions. Aesthetic emotions are elicited by aesthetic experiences, such as films, music, or books and audio books. Especially for music, aesthetic emotions are affected not only by the composer’s intended emotions, but by multiple facets like the musical interpretation, context, or the participant’s mood. This study investigated the research question whether (a) vocal delivery of a literary text (emotionally expressive vs. flat) and (b) mood states before listening to the text affect the frequency of aesthetic emotions. An online sample (N = 148) listened to either an emotionally expressive or flat version of the same audio-recorded text (a novel section describing a scene in WWI) and answered the PANAS (measuring mood states before listening to the text) and the Aesthemos (assessing aesthetic emotions after listening to the text). A MANOVA showed a main effect of vocal delivery, with higher frequencies particularly for fascination, beauty, being moved, and interest in the emotionally expressive condition, and boredom and ugliness in the flat condition. Additionally, negative (but not positive) mood states before listening to the text affected the frequency of aesthetic emotions, indicating that a previous negative mood state lead to higher frequencies particularly of confusion, uneasiness, surprise, insight, and nostalgia in the expressive condition. In conclusion, like with music, aesthetic emotions are not only a product of the content of a literary text, but also of contextual factors like the delivery and previous mood states.

ABSTRACT. Recently, it was suggested that negative affect motivates the adaptation of attentional control. According to this idea, van Steenbergen, Band, and Hommel (2009) argued that positive affect should counteract the motivation for control adaptation while negative affect should enhance the motivation. In line with their argument, van Steenbergen et al. (2009, 2012) found that positive gain feedback relative to negative loss feedback diminished the congruence sequence effect (CSE) which is an indicator for the adaptation of attentional control. In the present experiment, we attempted to replicate this valence effect of monetary gain/loss feedback on the CSE and furthermore, tested how motivational orientation (approach vs. avoidance) affects the effect of monetary feedback on the CSE. Ninety-five native German speakers (19-30 years old) were quasi-randomly assigned to one of two groups: One approach-oriented group were informed they would gain extra money after task completion, and one avoidance-oriented group were informed they would lose some money after task completion. The experimental procedure and materials were the same as that of van Steenbergen et al. (2009). Results demonstrated that negative affect boosts the CSE, however, in avoidance orientation only. This finding highlights the role of motivation for recent theorizing on emotion-driven control.

ABSTRACT. Several studies have shown that music interventions before and during surgery can lead to reduced anxiety and pain levels of the patient. The present study investigates the effect of a music intervention during the caesarean section on subjective (State Trait Anxiety Inventory, visual analogue scale) and objective (cortisol, amylase, heart rate, blood pressure) measures of anxiety and stress perceived before, during and after the caesarean on the day of surgery. The patients (N = 304) were randomly allocated to the experimental group, listening to music in the operating room, or to the control group, undergoing the procedure without music. The analysis revealed lower levels of subjective anxiety at the end of the surgery in the experimental group compared to controls. The objective parameters showed significant differences between the groups in salivary cortisol increase from admission to skin suture as well as systolic blood pressure and heart rate, indicate lower stress and anxiety levels in the music group. These results propose that listening to music during a caesarean section leads to a reduction of anxiety and stress experienced by the expectant mother. Music during caesarean is an easy implementable and effective way of reducing stress and anxiety of the expectant mother.

ABSTRACT. Media sources differ in attributed levels of credibility. But does the level of trust placed in different sources affect how the information impacts people’s opinions and judgments derived from news? In this pre-registered study, we investigate influences of source credibility on person perception, evaluation, and deliberate person judgments based on affective person-related information. Participants (N=30) read headlines about negative, positive, or relatively neutral social behavior (e.g. Berlin: This casino owner forced teenager to smuggle cocaine) that were presented simultaneously with unfamiliar faces in visual contexts relating to existing, well-known German online media sources associated with low or high levels of credibility. After learning, judgments of persons paired with negative information were more negative than those related to neutral information, independent of source credibility. Similarly, event-related potentials (ERPs) indexing perception-related motivated attention (early posterior negativity, EPN) were affected by the emotional contents. ERPs related to higher-order person evaluation (late positive potential, LPP) showed a modulation due to the implicit appraisal of source credibility. However, this modulation was relatively small compared to the strong impact of the affective information that was present for both, credible and less credible sources. These findings demonstrate that information from media sources of low credibility nevertheless strongly affects emotional responses and deliberate person judgments.

ABSTRACT. The current surprise literature suggests that surprise is about both expectation and explanation. In three experiments (total N = 679) we investigated whether the behavioral, affective, experiential and cognitive surprise components are actually driven by how strong an event deviates from what was expected and by how easily it can be integrated with the constraints of an activated schema. We therefore developed a new paradigm which allowed us to elicit different shades of surprise by manipulating both the degree of deviance and the constraints of the activated schema. Participants were instructed that they would see ten stimuli of a certain type on the screen. Crucially, we manipulated the degree of deviance of the last stimulus and presented a stimulus that deviated to either no, medium or high degree from that one expected due to the activated schema. Orthogonally to that, we varied the constraints of the activated schema and induced a schema which had either high, moderate, or no constraints. As our dependent variables, we assessed behavioral response delay and explicit ratings of liking, surprise, and expectancy. Our findings underline the complexity of the surprise response syndrome by showing that the different components are only weakly correlated and do not follow a uniform response pattern. Expanding the scope of previous accounts, the present results further suggest that surprise is not only about the absolute invalidation of prior expectancies, but also about the degree of deviance and the ease of integrating a surprising event with the constraints of an activated schema.

ABSTRACT. Through consistent use in everyday life, colours are likely to be associated with emotions: green should be congruent with positive valence, and red with negative valence. Our study tests for such a congruence effect and investigates, whether it arises automatically and in isolation, or is due to polarity correspondence (mapping of the colours to the valence poles). Employing a valence categorization paradigm, participants (n = 141) saw, in a single experimental block, positive and negative German words either in only red (red monochromatic block), only green (green monochromatic block), or in red as well as green font colour (mixed-colour block). Median correct reaction times showed a significant congruence effect (faster responses to congruent items, i.e., positive words in green, negative words in red). This was influenced by block type: The congruence effect was significantly stronger for mixed blocks than for monochromatic blocks. Therefore, our results show a green-positive and red-negative association, but only when polarities of both dimensions are present and suggest to the participants to map perceptual cues (colour) as positive vs. negative poles of the conceptual dimension (valence), thus lending support to the polarity correspondence principle.

ABSTRACT. In four studies, we modified a competitive reaction time aggression paradigm that included emotional feedback from the victim via video clips. Participants were provoked and given the opportunity to punish their opponent. The punishment was followed by a short video clip featuring three distinct emotional reactions of the opponent: (1) anger, (2) sadness, (3) pain (and neutral displays as controls). We compared the punishment intensity that was selected by the aggressor in the trials preceding and following the emotional victim feedback. We found a consistent effect of the facial pain display reducing aggressive behavior in all 4 studies (overall N= 192), supporting the suffering hypothesis. None of the facial displays benefited from a direct feedback (via scale) indicating pain, stressing the strength of employing natural pain displays in contrast to methods that rely on inferring pain (e.g. pain-o-meters). Pairing the facial display of pain with a direct feedback indicating anger did result in a descriptive decrease of the pain effect (dz pain high anger: 0.26; dz pain low anger: 0.43). We found no significant effect of the facial anger display on aggressive behavior. All differences after anger displays do not significantly differ from the neutral baseline, inconsistent with the understanding hypothesis. Facial expressions of sadness only reduced aggressive behavior if disambiguated by direct feedback indicating the absence of anger. This is especially interesting since the pre-rating indicated that sadness expressions are among the easiest to correctly classify.

ABSTRACT. The ability to flexibly switch among tasks is fundamental to ensure goal-directed behavior in a dynamic environment. However, this flexibility comes along with costs, such as the n-1 switch cost or the n-2 repetition cost. The n-1 switch cost refers to the performance slowing and increase in error rates for task switches relative to task repetitions when switching among two tasks. The n-2 repetition cost is observed when switching among three tasks and refers to the cost for n-2 task repetitions relative to n-2 task switches. Typically, the n-1 switch cost is assumed to result from reconfiguration processes or proactive interference from the previously activated task-set. The n-2 repetition cost is assumed to result from lingering inhibition that reduces competition among task sets in trial n-2. Whereas explanations for both costs are sometimes integrated, it is so far unclear whether both costs are related. To examine this, we decomposed the processes underlying both costs in three task switching experiments (Experiment 1: N = 16, Experiment 2: N = 24, Experiment 3: N = 24) using the diffusion model as well as the linear ballistic accumulator model. The results showed that n-1 switch cost reflects interference caused by the residual activation of the previous task set as indicated by slower evidence accumulation processes. In contrast, there were no consistent parameter modulations underlying n-2 repetition cost. These findings emphasize that different cognitive processes are involved in n-1 switch cost and n-2 repetition cost and question the role of inhibition for task switching.

ABSTRACT. This study investigated the processing mechanisms of dual-memory retrieval of two distinct task responses after practice. In this task situation, on the presentation of a color word cue, participants either performed a left/ right keypress or gave a verbal digit response; under dual-memory retrieval conditions, both responses were executed. Previous research in this situation demonstrated that dual-memory retrieval practice leads to learned retrieval parallelism in a subset of participants. According to the set-cue bottleneck model, these participants learned to chunk the two responses, such that both responses could be retrieved in one pass through a retrieval bottleneck. Further, that chunking process is only possible when both responses are concurrently in working memory. There exists however no direct empirical evidence for the contribution of the working memory component to this chunking process so far. As a consequence, we investigated whether the modulation of an additional load on working memory has an effect on the occurrence of learned parallelism. This load was modulated from low load of one digit in a first group of participants (N = 24) to high load of six digits in a second group of participants (N = 24). Consistent with our predictions, high working memory load in comparison to low working memory load reduced the occurrence of learned parallelism after dual-memory retrieval practice, providing empirical support for the contribution of the working memory component to the chunking process in the context of the set-cue bottleneck model.

ABSTRACT. So far, misinformation research has exclusively focused on the memorial impact of ONE misleading post-event detail relative to ONE original event detail (even if the same misinformation is sometimes repeated). In real-life situations, however, witnesses may encounter multiple different (and mutually contradictory) pieces of misinformation relative to an event detail (e.g. from multiple press reports). Under such circumstances, the misinformation effect may be enhanced (because MORE misinformation is presented) or reduced (because the mutual contradictions undermine the credibility of the misinformation), or both processes may operate simultaneously. In four studies(total N = 238)testing these ideas, we found evidence for each of these possibilities. The experiments followed an expanded misinformation paradigm in which participants first witnessed an event (a harmless traffic accident on video) and later read TWO witness statements pertaining to the event. In different conditions, participants received (a) no misinformation concerning a critical item (control), (b) a single piece of misinformation from one of the witness statements (standard) or (c) two pieces of misinformation from the two statements that contradicted each other (contradictory). Some studies also included a condition in which the two witness statements contained an identical piece of misinformation (consistent), and further used highlighting of statement information as a way of inducing discrepancy detection. Assignment of items to conditions and witness statements was counterbalanced. Discrepancy detection was assessed, and analyses relating discrepancy detection to the misinformation effect size in different conditions will be presented.

ABSTRACT. Previous research indicates that intentional forgetting might prevent overloading of a cognitive system by decreasing the accessibility of outdated information. Following this idea, outdated information should be overwritten to increase the functionality of the system. To investigate this process, pairs of synonyms (53 word pairs) were created. In an ambiguous situation, the accessibility of one item should be decreased if the other item of the pair was learned. This in turn should increase the time needed to decide between the two alternatives. Before encoding, participants (N 45, 15 per group) were instructed to (a) either remember both items of each pair, (b) forget one item from each pair and remember the other (item-method directed forgetting condition) or (c) received only one item from each word pair and one unrelated item and to remember both. Following the encoding phase, ambiguous fill-in-gap sentences were presented. Participants completed them using the learned information. Finally, a free recall test was included to investigate the accessibility of all items. Reaction times for fill-in-gap sentences were measured using eye tracking and voice onset time. Preliminary results show that participants in the directed forgetting group did neither perform worse in the fill-in-gap task nor recall less words compared to the other two groups during free recall. However, they produced significantly less false items during free recall and preliminary data analyses indicate that they performed the fastest in the fill-in-gap task. This pattern of results will be discussed in the context of adaptive intentional forgetting.

ABSTRACT. The main assumption of chunking theory is that knowledge about semantic units in a certain task domain can be used to compress incoming information. Although the maximal capacity of immediate memory is 4±1 information units, this information compression allows to enlarge the memory capacity by increasing the size of information units. As even random information can be compressed to a certain extent, a capacity limit of 7 single bits of information is typically found with random material. These assumptions were tested with a complex span task. Expert musicians (n=75) memorized the pitch of a single note and then played a simple melody on a piano at first sight. This procedure was repeated twelve times and then a memory test followed in which all memorized notes had to be recalled in correct order. The presence of semantic units was varied within-participants: in experimental trials, consecutive notes formed major chords, while in control trials they did not. As expected, mixed model analyses revealed that memory capacity was larger when the material contained semantic units. In experimental trials, the mean capacity limit was 3.18 chords while in control trials it was 7.32 single notes. The TBRS*C computational model integrates time-based decay, refreshing mechanisms and chunking. It is explicitly suited to model recall performance in complex span tasks. This model will be used to analyze the data in order to confirm its validity.

ABSTRACT. Task switching leads to reduced selectivity in subsequent memory for task-relevant (targets) over task-irrelevant (distractors) items. This effect is modulated by top-down cognitive control during task switching (Richter & Yeung, 2012; 2015). In order to disentangle the relative effects of cognitive control on targets and distractors, we conducted two experiments, in which we manipulated cognitive control during the encoding phase and tested memory for targets and distractors in a subsequent surprise recognition test. The two experiments differed only in the encoding phase where either random (exp. 1, N = 160) or predictable (exp. 2, N = 160) task switches were administered. In both experiments preparation duration (150 ms vs. 1200 ms) and stimulus duration (500 ms vs. until response) were varied between-subjects in order to investigate the effects of top-down and bottom-up control, respectively. We hypothesized that bottom-up control would also modulate memory selectivity, defined as the ratio of recognized targets to distractors. Generally, manipulations led to diverging effects on targets and distractors. Enhancing target memory was at the expense of distractor memory and vice versa. Longer preparation led to higher memory selectivity but only in the random task switching experiment while there was no effect of preparation in the predictable task switching experiment. Interestingly, in both experiments higher memory selectivity was obtained through longer stimulus presentation. These findings render strong support for a link between cognitive control and memory in the sense of a cognitive mechanism controlling the gate to long-term memory.

ABSTRACT. The goal of feedback evaluation is to enable learning, and thus improving future decision making. In contexts such as gambling, however, feedback cannot be used to improve performance due to its missing goal relevance, i.e., its informativeness to optimize subsequent choice behavior. In this study, we compared neural correlates of feedback processing, the feedback-related negativity (FRN) and the P3, in a probabilistic learning task and in a gambling task. Feedback sequence was yoked across tasks and participants were informed that feedback information can be used to improve performance in the learning task, but not in the gambling task. We hypothesized that only in the learning task should the neural correlates of feedback processing reflect behavioral adaptation. Our results provided support for this hypothesis as both components, particularly the P3, were modulated by task type, with only feedback activity in the learning task being sensitive to subsequent behavioral change. Likewise, computational modelling suggests that central parameters of a reinforcement learning model are related to feedback activity only in the learning task. Single-trial analyses revealed that in the learning task, but not in the gambling task, the FRN and the P3 reflected reward prediction errors derived from the reinforcement learning model. Together, these results point to the importance of goal relevance as a key feature of feedback processing.

ABSTRACT. Adaptive behavioral control involves a balance between top-down persistence and flexible updating of goals under changing demands. According to the metacontrol state model (MSM), this balance emerges from the interaction between the frontal and the striatal dopaminergic system. The Attentional Blink (AB) task has been argued to tap into the interaction between persistence and flexibility, as it reflects over-persistence—the too-exclusive allocation of attentional resources to the processing of the first of two consecutive targets. Notably, previous studies are inconclusive about the association between the AB and non-invasive proxies of dopamine including the spontaneous Eye Blink Rate (sEBR), which allegedly assesses striatal dopamine levels. We aimed to substantiate and extend previous attempts to predict individual sizes of the AB in two separate experiments with larger sample sizes (N=71 & N=65) by means of non-invasive behavioral and physiological proxies of dopamine (DA), such as sEBR and mood measures, which are likely to reflect striatal dopamine levels, and color discrimination, which has been argued to tap into the frontal dopamine levels. Our findings did not confirm the prediction that AB size covaries with sEBR, mood, or color discrimination. The implications of this inconsistency with previous observations are discussed.

ABSTRACT. The ability to monitor our actions and re-adjust our cognitive system when necessary enables us to continually improve our performance. In this process, response conflict detection is thought to play a key role. Due to the temporal resolution of event-related potentials (ERPs) which allow to disentangle cognitive processes in real-time, the error-related negativity (ERN) occurring immediately after an incorrect response has been evaluated as an index of response conflict detection. However, so far little is known about task-specific modulations within the same individual over time. Here, we investigated the extent to which performance and the underlying neural activity are modulated over time in different tasks in a student population (N = 21). Participants performed modified flanker and stroop tasks in two sessions one week apart while EEG was recorded. In the Flanker task, a target arrow was presented in the middle of the screen flanked by a total of 8 arrows. The Stroop task consisted of four different color words presented in either the same or a different color. Each task consisted of 160 congruent and 160 incongruent trials. Behavioral performance was largely stable over time, but showed considerable individual differences, in particular in terms of response accuracy. ERN amplitudes gradually diminished over time, but for each task only subtle differences between sessions were apparent. Notably, task specific effects on ERN amplitudes were evident, suggesting that partly dissociable cognitive processes underlie error-related brain activity. Together, the present results indicate task-dependent inter-individual differences, but intra-individual stability on error monitoring over time.

ABSTRACT. To deal with the limited capacity of the visual system, attention allows us to select relevant information from our environment and working memory allows us to maintain the selected information over time. Previous research has focused on the prioritisation of relevant information in attention and working memory and neglected the processing of irrelevant information, although there is a growing body of evidence showing that performance in visual tasks is strongly dependant on the efficiency to suppress irrelevant information. Here, I will present data from various studies that demonstrate the importance of suppression in visual selective attention and working memory. A powerful tool to investigate visual suppression is the distractor-positivity (PD), an event-related potential component in the EEG signal that is typically elicited contralateral to a salient distractor. It reflects active suppression of irrelevant information and is the counterpart of the N2pc component which reflects prioritisation of a stimulus. We used the PD component to provide evidence that suppression efficiency is affected by associate learning (n = 28) and by monetary rewards associated with the distractor (n = 27). We also used the PD component to show that active suppression plays an important role in in working memory processes (n = 46). Lateralised components like the PD and N2pc are limited to comparing attentional processes between hemispheres. However, with an inverted encoding models (IEM) that make use of multivariate signals we were able track the locus of attentional suppression at various location simultaneously (n = 36).

ABSTRACT. Imagine yourself driving quickly but safely while following your navigation system. When approaching a red traffic light, you can slow down. Moreover, if you see a police officer at the junction, you are able to ignore the traffic light and the direction signalled by the navigation system in order to follow the officer’s instructions. The goal of the present study was to investigate whether encountering such situations in which multiple conflicts are presented concurrently results in a simultaneous or sequential conflict resolution. To this end, we measured event-related potentials (ERPs) in a paradigm combining a Stroop and a flanker task. In this paradigm, colour words were printed in colour and the colour of the central letter was congruent or incongruent to the meaning of the word (Stroop task) and also congruent or incongruent to the colour of the flanking letters (flanker task). Twenty-four participants were asked to indicate the colour of the central letter while ignoring the meaning of the word and the colour of the flanking letters. The behavioural results replicated previous findings: The Stroop congruency effect (i.e., the difference between Stroop incongruent and congruent trials) was smaller – but still significant – for flanker incongruent than for flanker congruent trials. The ERP results showed that the flanker conflict was associated to an early ERP component (P2), whereas the Stroop conflict was associated to a later component (N450). Together, these findings emphasize a sequential organization of conflict resolution in the brain which is adaptive when facing multiple concurrent conflicts.

ABSTRACT. We tested if stimulus-response (SR) compatibility effects of spatially ambiguous words depend on a semantic priming context. Although many words, including spatial words, can take on several meanings, this is an open question. From Experiments 1 to 3, we manipulated the likelihood that the vertical meaning of the German particles auf and ab was processed by (1) instructing the processing of vertical meaning in Experiment 1, but not in Experiments 2 and 3, and (2) by using verbs that either primed (Experiments 1 and 2) or did not prime (Experiments 1–3) the targets’ vertical meanings. Spatial SR compatibility effects resulted, regardless of whether or not the processing of the vertical meaning was instructed and the vertical meaning was primed. Results suggest that the selection between vertically discriminated responses could be sufficient to elicit the participants’ extraction of the vertical meaning of the ambiguous particles.

ABSTRACT. Various features of words have been shown to impact attitudes towards words and the persons or objects they denote, as for instance prototypical and therefore familiar word patterns are perceived positively. Linguistic research has shown that syllables and therefore words are more likely to start with a consonant and end with a vowel. Therefore, words that follow these prototypical linguistic principles (starting with a consonant, ending with a vowel) should evoke positive attitudes. In ten (total N = 924) experiments, the impact of starting and ending letters on preference ratings and person perception was tested. Indeed, we found that for the starting letters of pseudo words, consonants were preferred over vowels. For the ending letters of pseudo words, vowels were preferred over consonants (the CV-effect). Words that started with a consonant and ended with a vowel were more likely to be rated as real existing words, which substantiates the notion that the prototypicality of CV-words underlies this preference effect. The effect was applied on preferences of person names, person perception and in an economic decision-making paradigm. It was replicably shown that persons and counterplayers with CV-conform names were liked more, were rated to be more competent and warmer and were allocated higher amounts of money. Implications of this novel language preference effect for the understanding of language processing and experimental set-ups are being discussed.

ABSTRACT. One reason for misinterpreted risks is inadequate risk communication among stakeholders. This works aims to address the key challenge of relative risks (e.g. 30% more X) with a tool that provides supporting information to the reader of digital texts: Explanations, if absolute risk information is missing because neither baseline risk nor effect size can be inferred; visuals, if absolute risk is reported because low numerated would benefit from them. However, it is not known whether enriching relative risk statements with transparent explanations (about missing information) and relative+absolute risk statements with transparent visuals improves readers' interpretation of risks. 215 laboratory participants read pretested modifications of German online texts about 13 different topics (e.g. sports, finance). The texts contained manipulated statements with single/multiple relative risks with/without absolute level (target conditions) or with single/multiple proportions in percentage/natural frequency format (control conditions). Four groups received plain text, explanations/fact box visuals (embedded or mouse-over), or the statements were simply highlighted (between-subjects). While under control conditions extraction and inference from the text was not systematically improved, embedded explanations and visuals improved both in each target condition. Statement-related risk perceptions were reduced only under the target conditions. Based on a wide range of variations on how to express a relative risk in words, provided explanations (which prompt critical statement review) and visuals (which convey part-to-whole relationships for low numerated individuals) can improve readers' interpretation of risk information. Applied research needs to show whether using this tool promotes learning or creates dependencies and hampers attention of readers.

ABSTRACT. Cognitive flexibility, that is a flexible adaption of processes ensuring goal achievement, is usually measured using the task-switching paradigm. In this paradigm, participants are requested to switch between two simple tasks. Performance costs arise whenever the currently relevant task is different from the task one trial before (trial n-1). Likewise, given a more and more multilingual world, we are often asked to switch among languages when addressing different colleagues in the office, for instance. Performance costs are found in this context, too. The question of whether performance costs arising in task switching can be attributed to the same processes as those in language switching is currently under debate. We addressed this question by fitting learning curves to task-switching and language-switching performance in three samples of bilinguals (Russian-German, Russian-Hebrew and Turkish-German) and their monolingual counterparts. We predicted different learning curves for bilinguals, suggesting more memory-based performance (i.e., retrieval of recent episodes) in language switching compared to task switching (i.e., engaging in task-set reconfiguration) whereas no such difference should arise in monolinguals. Such a pattern would suggest that individuals' language proficiency and use might influence the degree to which task switching and language switching recruit similar underlying processes.

ABSTRACT. Words with front-to-back consonantal articulation trajectories (inward, e.g., BAKA) are preferred over words with back-to-front trajectories (outward, e.g., KABA), an effect established by several independent labs recently. In this project the minimal required conditions for this in-out effect to occur are explored. In all previous studies whole words were shown, while here we show only word fragments to orthogonally manipulate starting and ending letters. In Experiment (N = 100) we show word fragments with front vs. back starting and ending letters (e.g., B_ K _, inward; K _ B _, outward) and still find the in-out effect. In Experiment (N = 120) we only present front vs. back starting and ending letters (e.g., B _ _ _, starting front; K _ _ _, starting back; _ _ B _, ending front; _ _ K _ ending back) and do not find the in-out effect. Experiment 3 (N = 200) corroborates this pattern with a between-subjects design. These studies show that the in-out effect does not hinge on the identity of the starting or ending letter per se but that the whole front-to-back vs. back-to-front trajectory is necessary to evoke it.

ABSTRACT. Lesen durch Schreiben (LdS; “reading through writing” in English) is a controversially discussed method for primary reading instruction in German-speaking countries. In LdS, primary school students learn reading and writing through prolonged “invented spelling”, meaning only phonological but not orthographic spelling errors are corrected. Previous studies of the effectiveness of LdS reported inconsistent results, casting doubt onto the suitability of LdS for primary school instruction. Furthermore, there are many methodological evaluation difficulties, such as separating the effects of LdS from other instruction-related variables.

In this work, we propose a connectionist model based on recurrent neural networks. The model enables us to test the LdS teaching regime in isolation, circumventing some of the previous challenges. We compared the behaviour of neural agents trained using the LdS regime against agents trained using a classical, primer-based training regime. Experimental results revealed that our LdS agents performed significantly worse than our primer agents in writing tasks and, to a lesser extent, in reading tasks. The LdS agents frequently repeated stereotypical spelling mistakes of children exposed to LdS. These mistakes arise naturally during writing acquisition of all learning agents, but are either suppressed or reinforced depending on the learning regime. We examined the learned, internal representations of both agents and found deviations in the LdS agent that may have induced the amplified confusion of similar phonemes. Our work presents a modest step towards building a biologically inspired framework of primary reading instructions that may act as a starting point for future research.

ABSTRACT. In French, Italian, and English double consonants have been shown to have a special status during the handwriting process. We investigated the role of double consonants in German handwriting, recording writing durations and pauses between letters during the handwriting process in four groups of participants: primary school children (3rd and 4th graders) (n=60), 6th graders (n=26), young adults (n=22) and adults (n=20). We expected that the planning of double consonants in words like TANNE (fir tree) differs from the planning of corresponding non-double consonant letter pairs, such as NT in TANTE (aunt). Such differences should be detectable in differences in stroke durations and inter-letter interval durations between the double consonant words and the control words. Participants copied the words that were presented on a screen onto a graphic tablet. Their writings were recorded using the software Ductus. Linear-mixed effect models with either stroke durations or inter-letter interval durations as dependent variables and position and word type as predictors revealed that double consonants had a different status in handwriting compared to non-double consonant letter pairs, detectable in significant main effects of word type or an interaction between word type and letter position. These effects differed between the groups of participants. Overall, results suggest that double consonants are planned together as one unit in German handwriting, which mirrors the results of studies in other languages.

ABSTRACT. Semantic interference during language production – can the effect be modulated? An eye tracking study

Experimental manipulations of semantic context during language production, for example through Picture-Word-Interference (PWI), have consistently shown that semantically related distractor words (e.g., cat) inhibit retrieval of the target picture name (dog). Here we tested the hypothesis that this effect may be parametrically modulated, i.e. that interference should increase with more distractor words presented. Using a new variation of the PWI paradigm we manipulated the number of distractors that could interfere with picture naming. In each trial, we presented 8 words in a circle, out of which 3 to 5 belonged to the same semantic category (e.g., horse, donkey, goat). Consecutively, participants (n = 24) named a picture that was either related to these context words (sheep), or unrelated (apple). We furthermore tracked their eye movements to investigate how they processed this semantic context. Our results show that the participants fixated longer on those words that were semantically related to each other. Overall, we replicated the interference effect for related vs. unrelated pictures. However, participants seemed to profit more from a bigger cohort of distractor words, because naming was faster with 4 or 5 than 3 distractor words. This effect appeared strongly only at the first naming instance of every picture and dissipated across the experiment. Our findings therefore indicate that besides interference during first exposure, repeated exposure to the semantic context, and a bigger lexical cohort (5 vs 3 words) may facilitate picture naming.

ABSTRACT. Lexical-semantic representations and processes in German-Turkish bilingual primary school children

A large proportion of primary school children in Germany grow up with German as their second language. They typically begin to acquire German when they join Kindergarten. Starting primary school, their German proficiency is often less advanced than that of their monolingual peers. Nevertheless, German is the teaching language at school, posing a challenge to many multilingual children. Using the blocked-cyclic naming paradigm in naming, reading, and translation tasks, we investigated behaviorally the lexical-semantic representations of 17 Turkish-German children aged 9 to 11. 19 German children served as a control group. As expected, there were no context effects on reading aloud in either group. In picture naming, Turkish-German children showed similar context effects as their monolingual peers in German. The corresponding effect in Turkish picture naming was much larger, as were the effects on translating words. While translating from German into Turkish caused similar effects as picture naming in Turkish, translating from Turkish into German yielded strongly exacerbated context effects. These results suggest that in Turkish-German speakers, lexical-semantic representations in German are similarly strong as those of their monolingual peers. The corresponding Turkish representations are less stable, possibly because the children use them less frequently than the corresponding German representations. This may be the case because the language dominance of the children has changed from Turkish to German in the first years of primary school.

ABSTRACT. The influence of socially shared information on semantic context effects

In dialogues, interlocutors usually establish mutually shared knowledge to enable successful conversation. Yet how shared information influences lexical-semantic processing remains largely unknown. In the semantic blocking paradigm, participants name repeatedly pictures that are either semantically related to each other (homogeneous blocks) or not (heterogenous blocks). Naming latencies in repetitions after the first presentation are longer in the homogeneous relative to the heterogeneous blocks, known as semantic interference. Similar patterns have been observed when the stimuli are from different semantic categories but are perceived as having a common theme. This study examines whether information shared between two task partners can create such a common theme. Before each naming session, participants watched a video of their task partner narrating a short story. In the subsequent naming session, blocks consisted of pictures either related or unrelated to the narrative. In our first experiment (N=32), the expected interference effects were found only when participants named pictures first in the heterogeneous block and then in the homogenous block, but not in the reversed order. This pattern was replicated in a second experiment (N=32). We conclude that the observed interference effects are susceptible to the time elapsed between narrative and picture naming. A control experiment (N=32) demonstrates that semantic interference is not experienced (in neither block order) when the narratives are unrelated to the picture stimuli. Together our experiments indicate that semantic relations are flexibly shaped by information shared between two task partners.

ABSTRACT. Joint picture naming: Is lexical access simulated on behalf of the task partner?

Speaking is a fundamentally social activity. Yet, surprisingly little is known about the neurocognitive processes underlying speech production in social interaction. Here we investigate in a joint action setting a well-established effect in speech production research, cumulative semantic interference, which arises when naming a series of semantically related pictures. In our experiments, two task partners take turns naming pictures of everyday objects. In three behavioral experiments we were able to show that naming latencies increase not only in response to the number of semantically related pictures speakers name themselves, but also in response to pictures named by their partner. This suggests that speakers represent their partner’s actions and pursue lexical access on their behalf. In two electrophysiological experiments we aimed to identify neural correlates of such simulated lexical access. As expected, in both experiments participants showed an increase in naming latency with each additional semantically related picture they named. Correspondingly, participants showed an increase in posterior positivity between 250-400ms, a time window typically associated with lexical access. However, unlike our previous experiments, our speakers did not appear to be influenced by their task partner’s picture naming. Accordingly, we found no electrophysiological evidence of lexical access on behalf of the partner. We conclude that speakers do not always represent their partner’s naming response and discuss possible factors that may have limited the participants’ evaluation of the task as a joint action.

ABSTRACT. Semantic picture-word interference in language mapping with transcranial magnetic stimulation

Language mapping by use of neuro-navigated transcranial magnetic stimulation (TMS) is applied in clinical research to identify language-related cortical regions. Recently, we proposed a high spatial resolution approach for more detailed mapping of cortical sub-areas such as Broca’s region. Moreover, we introduced the reaction time based picture-word interference paradigm into TMS language mapping to increase the objectivity of the method and to explore functional specificity with high temporal resolution. In a first study on 12 healthy participants, we employed a multimodal phonological picture-word interference paradigm. The inhibitory effects of TMS on language processing indeed reduced the behavioral phonological priming effect, which is characterized by accelerated naming responses to target pictures accompanied by phonologically related distractor words, specifically at stimulation sites overlapping with the probabilistic cytoarchitectonic area 44. Here, we adapt the picture-word interference paradigm to explore semantic specificity in Broca’s region. In a pilot study without TMS, we compared multimodal and unimodal presentation of auditory and visual distractor words, respectively, for efficiency of naming response acceleration. As a result, an unimodal semantic picture-word interference paradigm was chosen for the TMS study presented here. First results indicate a more anterior distribution of the semantic processing. Our results complement previous functional imaging data suggesting separate regions for semantic, syntactic, and phonological processing in an anterior-to-posterior direction along the inferior frontal gyrus, supporting a structure-function relationship in Broca’s region.

ABSTRACT. Semantic generalization of response-effect compatibility? In the response-effect compatibility (REC) paradigm, participants respond faster when the irrelevant effect presented after the response is compatible with the response than when it is incompatible. Compatibility hereby can refer to stimulus identity (e.g., seeing the number two after having responded by saying "two" is compatible, whereas seeing the number two after having responded by saying "eight" is incompatible) but also to a categorical decision like the spatial location (e.g., a right-sided effect following a right-hand response is compatible, whereas a right-sided effect following a left-hand response is incompatible). In our experiments, we were interested to see whether compatibility in the REC paradigm is also effective when it refers to semantic similarity. We describe data from five experiments in which semantic similarity was implemented either by using translation-equivalent words from two languages (e.g., dog - Hund) or by using a category-exemplar relation (e.g., dog – animal). Across all experiments, we observed a REC effect mainly in those conditions in which the identical stimulus was used as response and effect (i.e., monolingual, same word condition; e.g. hearing the word dog after having responded with saying "dog"). REC effects based on semantic similarity were weak and inconsistent across experiments. These data suggest that endogenous response priming due to the anticipation of an effect in the REC paradigm mainly relies on an anticipation of perceptual (in our case phonological) rather than semantic response features.