ABSTRACT. Revisiting the Contingency Sensitivity of Evaluative Conditioning: An Ecological Conceptualization

We reassessed the sensitivity of evaluative conditioning (EC) to contingencies under an ecological framework. In an EC procedure, a conditioned stimulus (CS) is paired with a positive or negative unconditioned stimulus (US) and as a result acquires US valence. Our ecological definition of contingency is dependent on two aspects: The conditional probability of a US being positive given a specific CS occurs and the probability of a US being positive given any other CS occurs. We assume that the more a stimulus is predictive of positive (negative) valence, the more positive (negative) should the evaluative shift be. This implies relativity and context sensitivity of EC. That is, for a CS to demonstrate a positive evaluative shift, it needs to be paired with positive USs with a higher probability than other CSs in the environment. We manipulate ecological contingencies by introducing CSs paired with different ratios of positive and negative USs and by manipulating the overall positivity (negativity) of the context. In three experiments (total N = 252), we found support for the impact of the ecological contingencies on evaluative shifts. Further analyses showed that the effects are mainly driven by the ratio of positive to negative pairings of a stimulus rather than the overall positivity (negativity) of the stimulus context. This holds true for objective as well as subjectively perceived values of CS-US contingencies, conditional probabilities and valence base rates. The implications for theoretical models of EC are discussed.

ABSTRACT. A single-process account of extinction and awareness in evaluative conditioning

Evaluative conditioning (EC) has been discussed as a learning process qualitatively distinct from classical conditioning, mainly because it was found to dissociate from expectations in extinction protocols, and because it has been reported to occur in the absence of awareness. It features prominently in dual-process models of attitude learning which propose that EC reflects an automatic learning process independent of awareness, resources, and goals. Contrasting this view, I present several lines of evidence that are consistent with a single-process perspective on EC. Regarding awareness, EC in the absence of awareness could be found under either subliminal (i.e., weak stimuli that are attended, e.g., with brief and masked presentation) or preconscious conditions (i.e., strong stimuli that are unattended, e.g. under incidental learning conditions). Contrasting dual-process predictions, several studies show that EC requires awareness of stimulus presentations, as well as memory for critical US properties, for EC to obtain. Regarding extinction, the dissociation arises at the retrieval or performance (rather than the learning) stage: It is due to differences between the evaluation and expectation measures and does not require dual learning processes or representations. A single-process memory-based-judgment framework of EC is briefly discussed.

ABSTRACT. Due process in dual process: Or, how hundreds of category learning studies might be wrong

Behavioral evidence for the COVIS dual-process model of category learning has been widely reported in over a hundred publications (Ashby & Valentin, 2016). It is generally accepted that the validity of such evidence depends on the accurate identification of individual participants' categorization strategies, a task that usually falls to Decision Bound analysis (Maddox & Ashby, 1993). Here, we examine the accuracy of this analysis in a series of model-recovery simulations. Implications for due process in the future evaluation of dual-process theories, including recommendations for future practice, are discussed.

ABSTRACT. How do I know what my theory predicts?

To get evidence for or against one's theory relative to the null hypothesis, one needs to know what it predicts. The amount of evidence can then be quantified by a Bayes factor. It is only when one has reasons for specifying a scale of effect that the level of evidence can be specified for no effect. In almost all papers I read people declare absence of an effect while having no rational grounds for doing so. So we need to specify what scale of effect our theory predicts. Specifying what one's theory predicts may not come naturally, but I show some ways of thinking about the problem, some simple heuristics that are often useful, including the room-to-move heuristic and the ratio-of-scales heuristic.

ABSTRACT. On the poor correlation between awareness and performance in implicit learning research

Several analytical approaches to assessing the unconscious character of learning rely on analyzing the relationship between awareness and performance. For instance, if participants’ performance in a learning task does not correlate with their awareness of the stimuli, it is typically assumed that learning must have been unconscious. What these analyses often overlook is that the statistical relationship between measures of learning and awareness is biased by their respective reliabilities. In the present paper, we present evidence that the measures used in a popular implicit learning paradigm, contextual cuing, are remarkably unreliable and that once reliability is taken into account, the correlation between performance and awareness is relatively large. These results suggest that correlational analyses should be restricted to highly reliable measures of learning and awareness.

ABSTRACT. A Bayesian Perspective on Correlation Analysis to Investigate Unconscious Mental Processes

As a method to investigate the scope of unconscious mental processes, researchers frequently obtain concurrent measures of task performance and stimulus awareness across participants. Even though both measures might be significantly greater than zero, the correlation between them might not, encouraging the inference that an unconscious process drives task performance. We highlight the pitfalls of this null-correlation approach with reference to a recent study by Salvador, Berkovitch, Vinckier, Cohen, Naccache, Dehaene, and Gaillard (2018), which reported a non-significant correlation between the extent to which memory was suppressed by a Think/No-Think cue and an index of cue awareness. In the Null Hypothesis Significance Testing (NHST) framework, it is inappropriate to interpret failure to reject the null hypothesis (i.e., correlation = 0) as evidence for the null. Instead, a Bayesian approach is needed to compare the extent to which the data provide evidence for the null versus the alternative (i.e., correlation > 0) hypothesis, while considering the often low reliabilities of the performance and awareness measures. When applied to the Salvador et al. data, such an approach indicates virtually no support for the claimed unconscious nature of participants’ memory suppression performance. Researchers are urged to employ Bayesian methods to analyze correlational data involving measures of performance and awareness rather than NHST methods—ideally while accounting for the unreliability of these measures.

ABSTRACT. Not all onsets are created equal: ERP evidence for a top-down control over attention capture by onset cues

For a long time, researchers assumed that visual abrupt-onset singletons (i.e., stimuli that suddenly appear) in the periphery of the visual field capture attention in a bottom-up way. The underlying assumption is intuitive: Suddenly appearing peripheral stimuli (that might even pose a threat or an obstacle) have not been foveated so far, such that it makes sense to register these stimuli for a potential upcoming eye movement. Although past research casted doubt on this assumption and suggested top-down control over this kind of capture, clear-cut evidence from a continuous measure of attention capture is lacking. Therefore, we conducted an N2pc (a continuous electrophysiological marker of attention capture). Our participants searched for a white peripheral target and reported its shape. The target was either a white abrupt-onset target, presented in isolation, or a white color target, presented simultaneously with three red distractors. Each target display was preceded by a single peripheral abrupt-onset cue either in red or white. We found that the cue-elicited N2pc was restricted to color matching white abrupt-onset cues, whereas color non-matching red abrupt-onset cues were suppressed. In two additional experiments, we replicated these results and ruled out possible confounds like luminance differences. Taken together, our results demonstrate that participants exert top-down control even over attention capture by abrupt onsets.

ABSTRACT. Crowding across depth depends on target or flanker depth

Crowding refers to the impaired recognition of a peripheral target stimulus by adjacent flankers. Studies showing that target recognition is facilitated when the target owns distinct features suggest that crowding comes from uncertainty in spatial selection. Indeed, crowding especially depends on the spatial arrangement of target and flanking stimuli. In two Experiments, we investigated crowding across depth, aiming at the question whether depth as a spatial feature of target or flankers can facilitate target recognition and reduce crowding. Real depth was presented by superimposing the displays of two orthogonal arranged screens via a half-transparent mirror. Adjustment of screen distances created a fixation depth and a defocused depth. Participants’ task was to identify the gap position of a target Landolt ring (eccentricity 2°) surrounded by two flankers (spacing 1°). In Experiment 1 (n=20) flankers were presented defocused, in front or behind the fixation depth, while the target was presented on the fixation depth. In Experiment 2 (n=24) targets were presented defocused, while flankers were presented on the fixation depth. Our results indicate a systematic difference in crowding effects between varying target and varying flanker depth, which might be interpreted as an attentional effect based on task relevance of stimuli. When defocusing the task irrelevant flankers, crowding increased with increasing flanker distance from the fixated depth. When defocusing the task relevant target, crowding effects were increased when the target was in front of fixation depth compared to on and behind the fixation depth.

ABSTRACT. Pupil dilation and microsaccade rate relate inversely in decision-making

Pupil dilation and microsaccades, small and very fast eye-movements occurring under fixation, have become two popular variables that allow investigating ongoing cognitive processing. Increasing activation is linked to an increase in pupil diameter, while microsaccade rate is commonly suppressed. Investigating top-down processes, both pupil diameter and microsaccade rate have been linked to changes in cognitive load, visual search, or emotional activation. Moreover, main neural correlates of pupil dilation (Locus Coeruleus) and microsaccade generation (Superior Colliculus) are reportedly linked and covary in their activation. Surprisingly however, they have scarcely been combined, although literature points towards a potential redundancy in research. Consequently, the question arises whether pupil dilation and microsaccade rate produce redundant results in a number of investigations. To test this notion, two experiments investigating binary decision-making were conducted, during which targets and distractors were presented in a balanced frequency. While pupil dilated more when fixating a target compared to a distractor, microsaccade rate was lower for target compared to distractor in both experiments. These results support the idea of an overlap between pupil and microsaccade results in eye tracking research. Future investigations may profit from this fact by choosing the respective preferable variable.

ABSTRACT. The eyes have it: Event conceptualization in German and Korean speakers

This study investigates the grammatical influence of language on memory and gaze behavior of spatial events in German and Korean native speakers. German and Korean, the two languages contrasted in this study, differ strikingly in how spatial events are described and processed. In German, the path of a spatial event is expressed in prepositions or particles, whereas in Korean it is expressed in the verb root. We tested German and Korean native speakers in a linguistic description task as well as in an eye tracking memory task. Our results show that both verbal (linguistic expressions) and nonverbal (memory performance, eye movements) behaviors are determined at least in part by language-specific grammar. While German speakers focus on the motion of an event, Korean speakers’ attention is drawn by the result of a spatial event.

ABSTRACT. Proactive effect monitoring for forced-choice and free-choice actions

When our actions yield predictable consequences in the environment, our eyes often already saccade towards the locations we expect these consequences to appear at. Such spontaneous anticipatory saccades occur based on bi-directional associations between action and effect formed by prior experience. That is, attention is guided by expectations derived from prior learning history. The resulting anticipatory saccades are thought to reflect a proactive effect monitoring process that prepares a later comparison of expected and actual effect. Here, we examined whether action mode (forced-choice vs. free-choice) affected such anticipatory saccades. Participants pressed a left/right key to predictably produce a visual effect on the left/right side. Action and visual effect were spatially compatible in one half of the experiment and spatially incompatible in the other half. Across and within experiments, we manipulated whether participants effect-generating actions were performed in a forced-choice or free-choice action mode. Both when participants performed forced-choice and free-choice actions, we observed anticipatory saccades towards the location of the future effect preceeding the respective effect´s presentation. Importantly, neither the frequency, nor latency or amplitude of these anticipatory saccades significantly differed between forced-choice and free choice conditions. We further ruled out an influence of target-effect associations. Overall, our findings suggest that forced-choice and free-choice action modes have comparable effects on attentional processes involved in the proactive monitoring of future action consequences.

ABSTRACT. Action control levels in the oculomotor domain: The case of free-choice saccades. Models of eye movement control distinguish between different control levels, ranging from automatic (bottom-up, stimulus-driven selection) and automatized (based on well-learned routines) to voluntary (top-down, goal-driven selection, e.g., based on instructions). However, one type of voluntary control has yet only been examined in the manual, not in the oculomotor domain, namely free-choice selection among targets that are equally attractive both from a bottom-up and top-down processing perspective. Here, we ask which features of targets (identity-related or location-related) are used to determine such oculomotor free-choice behavior. In two experiments, participants executed a saccade to one of four peripheral targets in three different choice conditions: free choice (unconstrained), constrained choice based on target identity (color), and constrained choice based on target location. The analysis of choice frequencies revealed that free-choice selection closely resembled constrained choice based on target location. The results suggest that free-choice oculomotor control is mainly guided by spatial (location-based) target characteristics. We explain these results by assuming that participants tend to avoid less parsimonious re-coding of target-identity representations into spatial codes, the latter being a necessary prerequisite to configure oculomotor commands. Based on these and other results, a model of oculomotor control levels will be developed.

ABSTRACT. No matter if it is in election battles, in personal relationships, or in fake news – lying affects us almost every day. Since lying is such an important moral category, it does not come as a surprise that there has been a lot of empirical research about lying, for example about how to detect lies, the time children start telling lies, and how often people lie in everyday life. Naturally, empirical researchers in the described areas all assume a particular concept of lying - and philosophers have argued a lot about the right definition of lying, especially which definition best captures people’s use and understanding of this concept. However, there is surprisingly little empirical research about the concept of lying and what it actually means to lie. Virtually all popular philosophical and linguistic definitions of lying are bound to declarative sentences and based on “what is said”. According to these narrow definitions, you cannot lie by asking a question or by implicating something false. We put these restrictions to the empirical test. In a series of 3 experiments, we presented a total of 500 participants with cases of deceiving questions (Experiment 1 & Experiment 2) and true assertions (Experiment 3) carrying deceiving implicatures. Here, we present our findings and provide first evidence that the lay concept of lying might be broader than commonly assumed in the philosophical and linguistic literature, i.e. participants’ concept of lying seems to include instances of asking a question and false implicatures as lying.

ABSTRACT. Obedience is a form of social influence in which a person yields to orders from an authority figure that are in conflict with her/his own values and norms (Milgram, 1974). The present research adapted a bug-killing paradigm (Martens et al., 2007) for the study of obedience. The study allegedly investigated how people feel when they destruct things (cover story). In one condition, the experimenter ordered participants to shredder living bugs in a manipulated coffee grinder. Most people condemn killing animals (including bugs), leading to an internal conflict with the order to kill. This obedience condition is compared with a control condition, in which participants killed bugs on their own free will (between subjects) and/or in which the experimenter ordered the destruction of non-living objects (e.g., coffee beans; within subjects comparison). In Experiment 1, 22 out of 23 (96%) participants killed the bugs in the obedience condition, while only 7 out of 22 (32%) followed the experimenter’s instruction in the free-will condition. In Experiment 2, 22 out 30 (73%) participants killed the bugs in the obedience condition, while only 12 out of 31 (39%) followed the experimenter’s instruction in the control condition. Please note that, participants in Experiment 1 (but not in Experiment 2) were additionally filmed by a video camera. In both Experiments, participants showed clear signs of distress in the obedience condition, suggesting that they experienced conflict with the order to kill the bugs.

ABSTRACT. Research on moral judgment often focuses on actions causing a single outcome. Many actions, however, have both an intended primary effect and unintended foreseen side effects. In two experiments we investigated how people morally evaluate such situations. In Experiment 1 we presented participants with cases in which an action causes a negative side effect (harming people, animals, or plants). While the side effect was held constant across conditions, we varied the number of entities that were helped as a primary effect. Moreover, we manipulated whether these entities were saved from a threat or whether the action merely led to a further improvement. We found that subjects aggregated the outcomes across multiple effects when asked about moral justification. With an identical negative side effect, the action was seen as more morally justified the more entities were helped. Also, saving has more impact on judgments of justification than mere improving. We additionally asked subjects to assess moral responsibility and found that agents were held less morally responsible for the same side effect when their action’s primary effect was an instance of saving as opposed to improving. In Experiment 2 we replicated these findings and further showed that subjects regarded an action as less morally justified and held the agent more morally responsible for a negative effect when the negative outcome was intended by the agent rather than the positive one. When morally evaluating actions with multiple effects, people take both the causal structure as well as agents’ mental states into account.

ABSTRACT. For millennia, humans cooperate successfully and there is little doubt that everything our species has achieved since is owed to our abilities to establish and maintain cooperation. In general, the evolutionary advantage of cooperation lies in the surplus it usually creates compared to human enterprises that are approached alone. At the same time, if people cooperate to create a surplus they will inevitably face the challenge to distribute what they created. If hunters cooperate to bring down big game, they must decide who takes which parts of the animal. If workers and employers form a firm, the revenue is divided into wages or profits. Not surprisingly, distributions are not always equal. The central question of this research is how equal vs. unequal distributions affect peoples’ willingness to cooperate. In a series of incentivized experiments with the public goods dilemma (total N > 1300), we orthogonally varied the marginal per capita returns from the public good and the symmetry of the payoff structure. Both, lower returns and asymmetric (i.e. structurally unequal) payoffs decreased the level of cooperation. Furthermore, mediation analyses suggest that asymmetric payoff structures decrease trust and fairness evaluations which in turn lead to lower levels of cooperation. In sum, this research suggests that people generally cooperate because the cake will be bigger. But if they divide it unequally, it may never be baked in the first place.

ABSTRACT. In a self-serving fashion, people often endorse more lenient moral standards for themselves than for others (i.e., moral hypocrisy). However, under some conditions, the reversed pattern in the form of moral hypercrisy may emerge as well. We argue that hypercrisy, that is, the other-serving endorsement of more lenient moral standards for others versus the self, is a prevalent phenomenon in close relationships. In two studies (N = 567), across two different cultures, participants judged various hypothetical transgressions within the relationship more leniently for their romantic partners (Study 1) and close friends (Study 2, preregistered) than for themselves. We further examined whether perceptions that one individual’s needs can only be fulfilled at the other’s expense (i.e., relationship-specific zero-sum beliefs) would moderate this effect. As predicted, the hypercrisy effect was significantly reduced for participants with high zero-sum beliefs. These results suggest that divergent moral standards for the self compared to others in the form of moral hypercrisy may be more prevalent than previously thought. Importantly, such other-serving morality may contribute to positive relationship dynamics. At the same time, the present findings underscore the detrimental consequences of zero-sum beliefs.

ABSTRACT. Formative assessment and feedback has become an increasingly used tool to enhance learning yet there is little evidence concerning the aspects of feedback information that students in higher education utilise when provided with various types of feedback. Another open issue regards whether different knowledge representations are acquired dependent on the feedback provided. Previous research indicated that the provision of feedback fosters the learning of abstract rules while no feedback enhances memorisation of item-specific information. Using an artificial language learning task we investigated the role of feedback timing on the learning process. The talk will describe a series of experiments demonstrating the importance of timely feedback delivery in various scenarios. Using receiver operating characteristics (ROC) the contribution of rule- and item-specific representations was assessed in these experiments. Results indicated that timely feedback led to a high proportion of rule-representations and a low proportion of item-based knowledge while for delayed feedback the opposite pattern was observed. These results suggest that timely feedback is a necessary prerequisite for successful learning of abstract knowledge that can be transferred to novel situations.

ABSTRACT. Memory performance is enhanced after testing compared to restudying (i.e. testing effect). Providing feedback in the form of presenting the full material again after the retrieval attempt (test-potentiated encoding - TPE) further increases memory retrieval. So far, the repeated presentation of the material was combined in some studies with positive or negative feedback depending on the initial retrieval attempt, thus combining the effects of TPE and performance feedback. This study aimed to evaluate whether explicit positive or negative feedback further enhances memory performance beyond the beneficial effect of TPE without feedback. Healthy German native speakers (N=40, 24 females, 20-30 yrs.) learned 210 weakly associated German cue-target word pairs by either restudying or testing (two repetition cycles) after a first exposure to the full material. Depending on the accuracy of the retrieval attempt, the tested word pairs were followed by positive or negative feedback (50%) or no feedback (50%). All trials were followed by another presentation of the full word pair (TPE). As baseline condition, 35 word pairs were only presented once. One day later, a cued recall test was performed. Results replicated the testing effect, showing considerably better memory performance for tested compared to restudied items after 24 hours. Furthermore, higher memory performance was found for tested items when participants received feedback compared to items without feedback. This effect was independent of feedback valence. These findings suggest a beneficial effect of feedback on later memory performance when learning is improved by retrieval practice.

ABSTRACT. People often take into consideration others' opinions and suggestions when they are reminding themselves general knowledge facts. The main aim of our experiments was to investigate how cues provided by external sources influence participants’ answers to general knowledge questions. It was also of interest whether people are able to discriminate between reliable and unreliable sources of information and, then, use them respectively. In Experiment 1, a list of difficult questions with answers was sequentially displayed. Participants’ task (N = 24) was to decide whether the answer to a given question is true or false. Some questions were presented with cue allegedly coming from one of the two previous participants. It turned out that conforming to the reliable social source improved participants’ test performance, whereas conforming to the unreliable source did not impair it. However, participants were equally inclined to incorporate cues from both sources, regardless of their accuracy. We hypothesised that better knowledge would help to effectively discriminate the sources. For this reason, in Experiment 2 (N = 60) the learning phase was added and participants studied the correct answers. In one group, participants read the questions with the answers, in the other group they generated an answer to each question or guessed it, what was followed by corrective feedback. Such experimental manipulation allowed to distinguish the effect of two different learning strategies on conformity behaviour and to check whether generation of answers in the learning phase would be beneficial for discriminating sources.

ABSTRACT. Frontal-midline (FM) theta activity (4-8 Hz) is proposed to reflect a mechanism for cognitive control, which is needed for working memory (WM) maintenance, manipulation, or interference resolution. Interestingly, modulation of FM theta activity via neurofeedback training demonstrated transfer to some but not all types of cognitive control. Therefore, the present study investigated whether FM theta neurofeedback training enhances performance and modulates underlying EEG characteristics in a mainly proactive control recruiting delayed match to sample (DMTS) task and a mainly reactive control engaging Stroop task. Moreover, we explored the duration of transfer over two posttests. Over the course of seven 30-minute neurofeedback training sessions, a FM theta training group (n = 17) exhibited a larger FM theta increase compared to an active control group (n = 18) who upregulated randomly chosen frequency bands. In a posttest performed 13 days after the last training session, the training group showed better WM performance measured in the DMTS task that was additionally predicted by FM theta increase during training as revealed by linear regression analyses. Contrarily, transfer to the Stroop task was not significant, suggesting that neurofeedback training might enhance primarily proactive and barely reactive control processes. In a posttest one day after training, group differences in both tasks were not significant. Surprisingly, training-induced improvements in WM performance were accompanied by attenuated FM theta activity, indicating less demands on cognitive control as a function of training. Together, the present findings show that neurofeedback training shows transfer to cognitive control that manifests late after training.

ABSTRACT. Recent research has suggested that sequential changes in the prospect of performance-contingent rewards may influence the balance between cognitive flexibility and stability: whereas constant high reward prospect seems to promote cognitive stability, increasing reward prospect has been shown to promote flexible behavior in voluntary task switching paradigms. Previous studies, however, confounded cognitive flexibility regarding voluntary task choices with control processes during task execution. We present two experiments (n = 30 each) to dissociate these two processes by means of a double registration procedure, in which task choice is registered prior to task execution. The data yielded clear evidence for reward-driven modulation of the flexibility-stability balance already at the level of task choices, with higher voluntary switch rates when reward prospect increased as compared to situations in which reward prospect remained high. These results thus confirm that the prospect of performance-contingent reward can indeed promote either cognitive stability or flexibility depending on the immediate reward history.

ABSTRACT. It is easier to select one of two competing task sets when anticipating a partner to perform the same task. Does this compatibility benefit depend on the partner performing the very same response as oneself or is it sufficient for him to select the same abstract task set but perform it on different stimuli and therefore produce different responses? Based on the theory of event coding, it was hypothesized that task sets are integrated into people’s event representations and therefore even abstract task set compatibility can affect subjects’ responses. In three experiments, 24 subjects either named pictures or read words, and following the subject’s response a partner responded by performing the same or the opposite task, using either the same or different stimuli. In Experiment 1, the partner used the same picture-word combinations as the subject and thus compatible trials implied a complete response imitation. Compatibility benefits were observed. In Experiment 2, the partner used his own stimulus set, therefore producing responses that differed from those of the subject in both compatible and incompatible trials. No compatibility benefits were found. As this result leaves the possibility that the tasks were too difficult, Experiment 3 replicated Experiment 2 with a smaller number of stimuli. This time, compatibility effects were found. These results indicate that a partner’s abstract task set can be represented and facilitate task set selection, but only when the tasks are not too demanding.

ABSTRACT. The Crosswise Model (CWM; Yu, Tian & Tang, 2008) is an indirect questioning technique designed to control for socially desirable responding in surveys on sensitive topics. The Crosswise Model is based on an experimental approach and provides demonstrably more valid prevalence estimates for sensitive attributes than conventional direct questions (Hoffmann, Diedenhofen, Verschuere, & Musch, 2015). It also elicits higher levels of trust and understanding among participants than other indirect questioning techniques (Hoffmann, Waubert de Puiseau, Schmidt, & Musch, 2017). The Extended Crosswise Model (ECWM; Heck, Hoffmann, & Moshagen, 2017) has recently been proposed as an advancement of the method. This model allows detecting several forms of response bias without loss in statistical efficiency. In a first practical application of the ECWM, we assessed the prevalence of islamophobia among a sample of German university students by asking 1508 students to answer a questionnaire in either a direct questioning or the ECWM format. As expected, the ECWM provided higher prevalence estimates for islamophobia than a direct question, indicating that the model successfully controlled for socially desirable responding. An assessment of model fit revealed that there was no specific response bias in favor of one of the two answering options, lending further support to the validity of the prevalence estimate obtained. Our results underline the importance of controlling for socially desirable responding, and make the ECWM appear as a promising approach for future surveys on sensitive topics.

ABSTRACT. This paper aims to improve the understanding of the value of cloud-based experiment management systems for cognitive research and education. Using Cognition Lab as an example will explain the difference between web-based experiment and cloud-based experiment management software. While web-based experiments use the internet to recruit large populations of anonymous participants at low costs and sometimes study the internet itself, cloud-based systems are offering cloud-based experiment management systems as a service to be used in or outside of a lab. Cloud-based services help to bridge between ‘in-lab’ settings and lab-external studies without having to switch technology. Design time, run-time, and data processing functions are all accessible via one central web console. This facilitates the management and execution of single lab studies as much as it helps to orchestrate multi-device and multi-lab study designs. The same tests can run on dedicated test stations in a lab, or on the participant’s device at home. To illustrate the use of cloud-based technology with named participants, two case studies performed with Cognition Lab are presented where the ability to reach participants at home.

ABSTRACT. Experimental psychologists need to decide on measurement methods and analysis strategies. The best choice is often difficult to intuit, and common practice or community consensus prevail as criteria. In this theoretical contribution, we propose to compare measurement methods by their ability to recover the effect of well-known manipulations in independent benchmark experiments. We term this property "retrodictive validity" and demonstrate that it extends classical measurement criteria such as validity and reliability. This approach allows selecting the least noisy among different valid methods for measuring/analysing a psychological construct of interest. Beyond methods comparison, there are many further applications: power analysis, cost-benefit analysis for study protocols, quality assurance, and training. Analysis methods based on computational models may be particularly beneficial for ensuring high retrodictive validity, for example psychophysiological modeling, sequential sampling models, and latent variable models such as structural equation models and item-response models. However, the framework is not limited to such methods and can be applied across many branches of psychology, including operationist approaches and theory-free decisions such as setting outlier criteria.

ABSTRACT. In within-subject designs, it is desirable that stimuli presented in the different experimental conditions are as similar as possible. In recognition memory, for example, word sets presented under different instructions should not differ substantially in overall familiarity; differences in responses should be attributable to the experimental manipulation and not to the material. However, the creation of equivalent stimulus sets presents researchers with a difficult problem, especially when experiments employ many stimuli. We present a method that facilitates and automates this process. It is based on the theory of cluster analysis and can be applied whenever researchers have access to quantitative data describing the properties of their stimuli (e.g., the length of word stimuli). Usually, clustering methods are used to create sets of elements that are as different as possible; we show that the objective of creating similar stimulus sets can be formulated as an "anticlustering" problem that reverses the objective of classical clustering methods. To be widely applicable by researchers in psychology, we implemented anticlustering as a free and easy-to-use software package ("anticlust") for the statistical programming language R. A case study on the Chicago face data base illustrates the usage of the package. Tests on simulated as well as real data show that the anticlustering method produces satisfying results, obtained in a completely automated manner.

ABSTRACT. Social desirability biases constitute a pervasive problem in surveys assessing sensitive attributes as they typically lead to underestimation of the true prevalence. To overcome these biases, the Randomized Response Technique (RRT) increases individuals’ anonymity by adding random noise to each answer. It thus encourages respondents to answer truthfully and has been repeatedly shown to produce better prevalence estimates than direct-questioning formats. Due to the random noise, however, traditional RRT analysis is associated with low statistical power, resulting in large required sample sizes when aiming to detect small prevalence rates. As a remedy, we introduce a design that combines the RRT with Wald’s Sequential Probability Ratio Test (SPRT). In contrast to traditional analysis, sequential statistical procedures continuously monitor the data during the sampling process and terminate when a predefined criterion is met. Sequential analysis may thus substantially reduce the required sample size without increasing long-term error rates. We show how to implement the SPRT for common RRT variants in standard statistical software. Moreover, we demonstrate analytically and by means of simulations that on average it requires approximately 50% smaller samples than traditional analyses. Finally, we illustrate the efficiency of the proposed sequential RRT with an empirical example.

ABSTRACT. The assessment of people’s risk-taking propensity is often carried out by means of self-report measures. To date it is unclear how people arrive at their ratings. We present a series of studies aimed at closing this gap. Based on the assumption that people retrieve information from their autobiographical memory to form their judgments, we use “aspect listing” as a process-tracing tool. We model the influence of the number of aspects listed in favor of taking risks, and against taking risks; the order of aspects; and the strength of evidence of these aspects. In a first study (N = 50), we explored basic properties of the aspect-retrieval process and how it relates to participants’ self-reported risk-taking propensity. We found that the evidence strength of the first three aspects listed each had a high correlation with the risk-taking propensity rating (.38 ≤ r ≤ .51), whereas the aspects listed later each had a correlation around zero, suggesting that aspects retrieved early may likely be drivers of the ratings. Based on this information we conducted systematic simulation analyses to compare several cognitive models, which we test in a preregistered second study (N = 250). Finally, a preregistered third study (N = 250) consists of a retest after two months, in order to quantify the extent to which the stability of self-reported risk-taking propensity hinges on the stability of the retrieved aspects. Our results shed light onto the cognitive processes of self-reported risk-taking propensity and thus have implications for the interpretation of these measures.

ABSTRACT. To which degree is imitation facilitated by a focus on others? It is known that individuals imitate a wide range of different behaviors. Different theories suggest that factors related to a focus on others, as compared to a focus on the self, facilitates imitative behavior. However, this claim has never been systematically investigated. In the present research, we fill this gap and test the basic assumption within three steps. First, we tested whether personality traits related to self-other focus influence imitative behavior. To this end, we ran a high-powered study and assessed imitative behavior with two different tasks. That is, we assessed mimicry by means of action observation (Chartrand & Bargh, 1999) and automatic imitation with the imitation-inhibition task (Brass et al., 2002)—a reaction time based task that measures imitation on a trial-by-trial basis. In addition, we measured participants’ self-construal, individualism and collectivism, as well as need to belong as measures of self-other focus. The results indicate that none of the scales correlates with any of the assessed imitative behavior. Second, we ran two additional studies in which we experimentally manipulated self-other focus. These studies reveal partial support for the idea that a focus on others indeed facilitates imitation in the imitation-inhibition task. Third, we ran a meta-analysis on published as well as unpublished studies and tested the influence of self-other focus on imitative behavior. The overall effect supports the basic idea that a focus on others facilitates imitative behavior.

ABSTRACT. On most occasions, we interact with other individuals and do not act in isolation. However, cognitive research has only recently began to transfer individual experimental task setups to socialized versions of traditional cognitive paradigms. In this talk, we provide a series of experiments aimed at testing social and non-social moderators of the Joint Simon effect using a social variant of the standard two-choice Simon task (n=32). Two participants shared a visual Simon task, so that each person basically performed complementary parts of the task, which transfers the paradigm into a go/no-go Simon task for each person. Before running this Joint Simon task, we set both participants either in a cooperative or a competitive state by means of a dyadic game, a manipulation aimed at testing a possible transfer of cognitive processing states across two different tasks. We found significant Joint Simon effects for participants that were in a cooperative state and for participants that were in a competitive state. The Joint Simon effect for participants being in a cooperative state was significantly larger than for participants being in a competitive state. These findings suggest a transfer of cognitive states from one shared task to another shared task. We discuss the findings in terms of domain-specific and domain-general accounts of the Joint Simon effect.

ABSTRACT. Our life is full of situations where we work together for a shared goal. Recent research has addressed this by splitting up established cognitive conflict paradigms (e.g. Simon task) between a pair of co-actors in such a way that each participant responds to only half of the trials. However, one other source of conflict arises from the differentiation between the responding agents, i.e. is it you or me?. Therefore, we established a variant of a Go/NoGo paradigm with a simple color differentiation task between a pair of co-actors using task-irrelevant faces (one for each co-actor, stranger’s face) as background stimuli. In a forerunner study (Baess & Prinz, 2017), we have reported an own-face advantage for one’s own face, which was modulated by the joint Go/NoGo task setting and familiarity of the co-actors. Here, we will present the results of 2 experiments combing both, task-irrelevant agent correspondence and spatial correspondence (Simon task). Background faces with the relevant color dot were shown on the left or right side of the screen. Two different sources for correspondence, i.e. agent-identity and spatial stimulus-response compatibility were assessed in a joint and individual Go/NoGo task setting. An agent-identity correspondence effect was yielded for the joint as well as the individual Go/NoGo task setting, albeit more pronounced in the joint Go/NoGo task setting. However, the spatial correspondence effect was only yielded in the joint Go/NoGo task setting. Results are discussed in regard to relevance of two sources of correspondence in joint action.

ABSTRACT. It takes two to tango: Testing social moderators in two variants of the observational stimulus-response binding task. Observing how someone responds to a stimulus results in incidental stimulus-response (SR) bindings. These observationally acquired SR bindings can be retrieved on a later occasion and thus will bias current response tendencies towards re-execution of the observed response. I will present and review evidence from two experimental approaches to investigate social moderators on retrieval of observationally acquired SR bindings. The first approach, a dyadic interaction paradigm in which two co-actors respond in alternating fashion, indicates that retrieval of observationally acquired SR bindings are modulated by social relevance. For instance, they were contingent on some form of chronic (romantic relationships) or situational (cooperation/competition) interdependence between co-actors. Further evidence supports that perceived similarity between co-actors is an underlying mechanism which mediates retrieval effects in the dyadic paradigm. The second approach is a video-based version in which responses are observed on screen. This approach appears to be immune to social moderators: Manipulation of visual perspective, similarity, partner salience, partner relevance, and group membership had no modulatory effect on retrieval of observationally acquired SR bindings in the video-based paradigm. Together, these findings suggest that only the dyadic paradigm produces genuine “social” effects that can be modulated with appropriate manipulations. Implications for the suitability of both paradigms to study observational learning and joint action phenomena are discussed.

ABSTRACT. Does group membership modulate automatic imitation? A minimal group approach

A prominent idea in the literature is that we are more inclined to imitate people that are more like us. However, while it is well-established that bottom-up similarity is an important modulator of automatic imitation, the contribution of top-down similarity is less clear. One potential reason for this is that research on top-down similarity has so far been restricted to situations in which we do not have to choose who we imitate. Therefore, in three experiments, we tested whether social group membership influences who we imitate in situations with one in-group and one out-group member. More specifically, we assigned participants to a “blue” or “green” group based on their political orientation and tested whether this modulated automatic imitation of simultaneously presented blue and green hands. Across the three experiments (N = 39; N = 39; N = 66), we found no evidence that group membership modulated automatic imitation, neither when the hands made different movements (Experiment 1), nor when they made identical movements (Experiment 2 and 3). This was true even though explicit (Experiment 1) and implicit (Experiment 2) measures both indicated that participants were positively biased towards their in-group. Overall, these results indicate that group membership does not determine who we imitate in situations with both in- and out-group members.

ABSTRACT. Mousetrap-Web: Mouse-tracking in the browser

Mouse-tracking is a versatile method for monitoring the development of cognitive processes over time, particularly the temporal development of choices and the degree of conflict between response options. So far, this method has been limited to laboratory-based software, and not easily been available to researchers looking to conduct studies online. Thus, researchers forgo the advantages that internet-based research offers, such as the quick and efficient collection of larger and more diverse samples. As a solution, we introduce the mousetrap plugin for lab.js, a free online study builder. It provides a graphical interface for constructing experiments without requiring programming skills, and allows for the easy implementation of mouse-tracking studies. Mousetrap-web also integrates with the mousetrap R package for processing, analysis and visualization of the collected data. We also discuss the methodological challenges of moving beyond the laboratory and collecting mouse-tracking data in self-administered online studies, where the indiviual participants' hardware and environment are not as easily controlled. Our software is available free of charge as open source from https://github.com/felixhenninger/mousetrap-web.

ABSTRACT. Mousetrap: Open-source tools for advanced analyses of mouse-tracking data

The recording and analysis of mouse movements has become a popular method in psychological research to investigate the temporal development of preferences and the competition between response alternatives. In a typical mouse-tracking study, participants decide between different options that are presented as buttons on a screen. During this task, their cursor movements are continuously recorded. In this contribution, we present novel methods for analyzing mouse-tracking data as a means to test psychological theories. These include tools for the spatial clustering of trajectories, the mapping of trajectories onto prototypes, and assessing the temporal order with which different areas of interest were visited in a trial. In addition, we offer tools for visualizing movement trajectories via animations and heatmaps. These analyses and visualizations allow for answering different research questions, such as which option was initially preferred in a trial and determining if and how often participants changed their mind. In addition, they can be used to determine if movement trajectories are homogenously distributed across trials or whether different types of movements are present in the data. This question is a core issue in many mouse-tracking studies that use trajectory shapes to determine whether dynamic models (assuming a continuous competition in all trials) or dual-process models (assuming a mix of trials with little and high conflict) better describe a decision process in a specific task. All methods are implemented in the mousetrap R package, which is freely available from http://pascalkieslich.github.io/mousetrap/.

ABSTRACT. An application of mouse-tracking to psychopathology research: Impression formation in Borderline Personality Disorder

Individuals with Borderline Personality Disorder (BPD) show negative first impressions of others based on self-report measures, which is thought to reflect cognitive schemas that others are threatening. We hypothesized that this tendency for negative evaluations of others should also be reflected in participants’ computer mouse movements. We expected that negative evaluations of others should entail relatively straight trajectories, whereas trajectories for positive evaluations should be more curved or show changes of mind. The material included videos of 52 target individuals who briefly spoke about their personal preferences. We presented these videos to a group of BPD raters (n=36), a healthy control group (n=36), and a social anxiety control group (n=30). After viewing each video, raters were presented with eight adjectives (positive and negative, counterbalanced) and selected whether an adjective applied to the target person by clicking a ‘yes’ or ‘no’ button while their mouse movements were tracked. Contrary to our expectations, individuals with BPD did not show more conflict in their mouse movements specifically when making positive choices, but showed more conflict in all their decisions. This was reflected by greater maximum absolute deviations for all trials for the BPD versus the two control groups and by a greater overall percentage of change of mind trajectories when using a prototype approach to analyze the data. Results can be interpreted as corroborating studies that suggest BPD individuals consistently self-report lower confidence in decision tasks.

ABSTRACT. Modeling Multiple Process-Tracing Metrics in Risky Decisions Involving On-line and Off-line Information Acquisition

Process-tracing metrics, which include eye and movement tracking, are valuable for characterizing cognitive processes that cannot be captured using outcome-focused metrics. Eye-tracking fixations indicate attention and information acquisition over time, and movement-tracking has been leveraged in decision-making research to model cognitive conflict, evidence accumulation, and preference reversals. However, there has been limited research collecting both of these metrics simultaneously (Koop & Johnson, 2013), and even less work modeling these multiple metrics on a common framework to examine their interactions and interdependencies over time (Blaha, et al., 2017). Our research provided evidence that simultaneous eye and movement tracking data provide insights into attention fluctuation and perceptual updating over the course of trials and the embodiment of this shifting attentional focus in hand and arm movements. We collected concurrent movement and eye tracking data during a risky decision-making task where participants received either congruent or contradictory choice recommendations on stock options. Outcome, risk, and recommended choice were either presented on-line during the trial, off-line prior to the trial, or both, between blocks. This allowed us to separate evidence accumulation and evaluation from the decision phase more explicitly. We modeled the differences in trajectories depending on whether information was evaluated in a perceptual encoding phase separated from movement versus evaluation during movement execution. Eye-tracking provided additional insights into perceptual updating of movements by revealing which information participants attended to during each trial, and the interdynamics of eye and movement-tracking provided greater insight into action updating based on continuous perceptual evaluation.

ABSTRACT. Tracking changes of mind through mouse movements

The recording and analysis of computer mouse movements provide us with a window through which we can observe the dynamics of decision-making. In the current research project, we assess the potential of several analysis techniques of mouse-tracking data that aim to detect how often people change their mind in choice tasks. This includes a number of established mouse-tracking indices, such as the number of movement reversals along the horizontal axis. We also provide several newly developed methods for detecting changes of minds (CoMs), including one method based on dividing the screen into different areas of interest, one method that explores the movements along the horizontal axis in more detail and one method that assigns trajectories to different prototypes based on their similarity. To evaluate the methods, we conducted two mouse-tracking experiments that included manipulations to experimentally induce CoMs in one of two conditions. We examined how sensitive each of the CoM measures was to the experimental manipulations. In addition, we assessed the extent to which each method agrees with human raters who determined the number of CoMs by visually inspecting each movement trajectory. Results suggest the new methods outperform the established indices but that depending on the task under investigation and the criterion for evaluation (agreement with human raters vs. sensitivity to experimental manipulation) different methods are superior. We will discuss when each method can be employed to gain information about decision-making processes.

ABSTRACT. Title: The forward testing effect is reliable and unrelated to working memory capacity

Abstract: The forward testing effect refers to the finding that retrieval practice of previously studied information enhances learning and retention of subsequently studied other information. The present research questions were whether the forward testing effect (a) is related to student learners’ working memory capacity and (b) is reliable in a test-retest design. Two experiments were conducted (Experiment 1: n=240, Experiment 2: n=64). Participants studied three lists of items in anticipation of a final cumulative recall test. In the testing condition, participants were tested immediately on lists 1 and 2 after initial study, whereas in the restudy condition, they restudied lists 1 and 2. In both conditions, participants were tested immediately on list 3. The results of both experiments showed a forward testing effect, with interim testing of lists 1 and 2 enhancing list 3 recall. Additionally, Experiment 1 showed that the forward effect is unrelated to working memory capacity, whereas Experiment 2 showed that the effect is correlated between two experimental sessions using different item materials. Together, these findings suggest that the forward testing effect is both reliable and robust.

ABSTRACT. Reset of encoding contributes to the forward effect of testing

The forward effect of testing refers to the finding that recall testing of previously learned material can promote learning of new material. The present study reports the results of two experiments (n = 120 in both experiments), which examined whether the size of the forward effect varies with the new material's serial learning position. In each experiment, participants studied three lists of unrelated items. They were tested immediately on lists 1 and 2 (testing condition), or restudied the same lists of items (restudy condition). In both conditions, participants were tested immediately on list 3. Results of both experiments show enhanced list-3 recall in the testing condition, relative to the restudy condition, thus replicating the basic forward effect of testing. More important, the forward effect was more pronounced for early than middle and late list-3 items, and serial position curves in the testing condition were largely identical across lists 1, 2, and 3. The findings are consistent with the view that testing induces a reset of encoding, making encoding of the newly studied list as effective as encoding of the initially studied lists.

192 words

ABSTRACT. Metacognitive judgements can modify future learning: The role of covert Retrieval

Judging how likely a learned item will be recalled in the future alters one's memory for this particular information. These Judgements of Learning (JOLs) seem to enhance memory and eventually induce covert retrieval attempts under specific conditions, in particular when JOLs are solicited in response to only a part of the learned information, compared to the full information. Here we investigated whether such metacognitive judgements modify future learning of new materials, compared to a restudy and a testing condition. Participants studied five lists of 20 German nouns in anticipation of a final cumulative recall test. After the presentation of the first four lists, participants either restudied the list, made JOLs based on complete items, made JOLs based on incomplete items (i.e., items' first 2−3 letters), or they were tested on incomplete items. After learning List 5, participants were instructed to freely recall only this last list. Results demonstrated that, compared to restudy, testing and making JOLs based on incomplete items potentiated recall performance of List 5 and attenuated the number of intrusions from Lists 1−4 to a similar magnitude. In contrast, compared to restudy, JOLs based on complete items did not reveal any effects on future recall of List 5 or the amount of prior list intrusions. The present study indicates that, similar to overt testing, metacognitive judgements can reduce proactive interference, and thereby enhance future learning, to the extent that they elicit covert retrieval attempts.

ABSTRACT. Spatial Learning in Children: Testing Enhances Subsequent Learning of Spatial Information in Younger and Older Primary School Children

Research with adults has shown that testing previously studied information can enhance learning and retention of subsequently studied new information. Here, we examined whether this forward testing effect (FTE) occurs in children’s spatial learning. Younger (n = 28; 7-8 years) and older (n = 28; 9-10 years) primary school children studied four successively presented 3 × 3 arrays, each composed of the same nine objects. The children were asked to memorize the locations of the objects which differed across the four arrays. Following presentation of each of the first three arrays, memory for the arrays’ object locations was tested using a reconstruction task (testing condition), or the arrays were re-presented for additional study (restudy condition). Array 4 was presented and tested in both the testing and the restudy condition. Results revealed superior location memory for Array 4 in the testing condition, relative to the restudy condition. In addition, participants in the testing condition also made fewer confusion errors during reconstruction of Array 4; that is, they tended to misplace objects less to locations where they had been presented previously, indicating that testing reduced the build-up of proactive interference. Importantly, both effects were similar in size in the two age groups, suggesting developmental invariance of the FTE in the age range considered. The findings highlight the importance of testing for enhancing young children’s learning and retention of spatial information.

ABSTRACT. The role of error magnitude in errorful generation effects

Incorrectly guessing the English translations of foreign vocabulary produces better recognition of the correct translations than an equivalent period of study time. Three experiments examined whether this errorful generation effect is driven by the magnitude of the error. In each experiment, participants studied the English translations of Finnish words. On Generate trials, the participants first guessed the category (four-footed animal or item of clothing), before guessing the translation. Within-category errors were defined as trials in which participants guessed the correct category but wrong translation, while cross-category errors were trials in which participants guessed both the wrong category and translation. On Read trials, participants simply studied the intact word pairs for the trial duration. In Experiment 1 (N = 72), within-category errors produced better recognition of the correct translations than either cross-category errors or Read trials. Experiment 2 (N = 88) replicated this target recognition effect, but showed that it does not extend to associative recognition. In Experiment 3 (N = 46), self-reported recognition judgements were partially aligned with target recognition; participants correctly gave higher recognition judgements for within-category errors than cross-category errors. However, Read targets received the highest recognition judgements even though they were recognised most poorly. These results speak against error magnitude accounts of errorful generation effects, where larger (cross-category) errors would be expected to produce better target recognition than smaller (within-category) errors.

ABSTRACT. Testing your memory: Current research on the forward testing effect and the benefits of unsuccessful retrieval (6)

Plenum discussion.

ABSTRACT. To exploit the potential of vehicle automation and ensure broad public acceptance, driving comfort is considered a key issue. The research aim of the KomfoPilot project at Chemnitz University of Technology was to identify psychophysiological reactions to uncomfortable automated driving situations, which could serve as a basis for detecting and subsequently reducing discomfort in real-time during driving. A total of 40 participants, aged between 25 and 84 years, experienced three highly automated trips in a driving simulator, including six potentially uncomfortable situations per trip. Perceived discomfort was reported continuously using a handset control. Heart Rate (HR) and Skin Conductance Level (SCL) were assessed by the smartband Microsoft Band 2; pupil diameter and eye blinks were measured using the SMI Eye Tracking Glasses 2. Results showed specific reactions during situations that provoked moderate to high discomfort. The eye blink rate decreased and pupil diameter increased in uncomfortable situations that were visually monitored. SCL did not show specific effects, which, however could partly be attributed to measurement deficiencies of the smartband. HR decreased consistently during uncomfortable situations, which could be related to the phenomenon “preparation for action”. No psychophysiological reactions were observable for longer lasting and slowly evolving situations with lower reported discomfort. We conclude that situations provoking moderate to high discomfort elicit specific measurable psychophysiological reactions. The results serve as a basis for developing a real-time discomfort detection algorithm and will be validated in a subsequent driving simulator study as well as on-road.

ABSTRACT. So far, emotional spatial cueing studies used either auditory (Zimmer et al., 2015; 2016) or facial cues (Liu et al., 2015) and indicated independent of modality that fearful/angry stimuli attract our spatial attention as mirrored in P3-activity. Further, non-emotional multisensory studies indicated enhanced spatial cueing effects with multisensory versus unisensory cues in fronto-parietal networks (Mastroberardino et al., 2015). Therefore, we asked if such multisensory effects could be transferred to emotional presence in facial-sound combinations. In an ERP-study, a left- or right sided facial-sound cue preceded a visual target (white triangle) on the same (valid) or opposite (invalid) side. The facial-sound cues consisted of either two matching stimuli (either both fearful or both neutral), or a mixed combination, e.g. a neutral face with fearful sound and vice versa. Twenty-five participants ignored the facial-sound cue and signaled the direction of the triangle (up/down) as fast and accurately as possible with a button press. Behavioral analysis showed that responses were faster for valid than invalid trials after matching fearful combinations but inverted in neutral combinations, whereas there was no validity difference in the one emotion combinations. ERP-data showed an early P3a-activity that increased for invalid targets, specifically after cues with two fearful parts or at least the fearful sound. In contrast, a later P3b-activity increased for valid targets as revealed by a main effect. Therefore, at early processing stages, the amount of emotional presence modulated the spatial P3a-effects whereas at later temporal stages, emotional influence on spatial processing rather decreased.

ABSTRACT. The experience of negative emotions often indicates that something is wrong and thus a focused approach is needed to solve the problem. However, a recent study by Gable and Harmon-Jones (2010) showed that sad images had a broadening effect on attention in a global-local letter task. They argued that sadness is associated with a failure to obtain goals, which broadens attention to aid the consideration of alternative options. Yet their findings are at odds with other studies showing that sadness either had no effect or even the opposite effect on attention. We will first report the results of a cross-cultural replication study that was aimed at replicating this broadening effect on attention. Results showed no effect on the global/local processing bias, neither by the sad pictures nor by the sad videos (Experiment 1 and 2). Results from a study using an Eastern-Asian sample further showed that this finding is not modulated by a culturally related processing bias (Experiment 3 and 4). We will then present a second study looking at how induced sadness (Experiment 5) or moderate depression (Experiment 6) influences the three functions of attention: alerting, orienting, and executive control. Results showed no effects on alerting or on orienting, but participants who were sad or moderately depressed showed less flanker interference (i.e., increased executive control) compared to participants who were neither sad nor depressed. Overall, these results suggest that sadness can affects attentional control, but more at the level of executive control than at the level of spatial processing.

ABSTRACT. Perspective-taking is an empathic process that helps humans to better understand the thoughts and feelings of their conspecifics. It is often argued that this greater understanding of other people’s thoughts and feelings then contributes to more positive interpersonal relations. However, it is unclear why a better understanding should always result in beneficial social outcomes. Here, we investigated whether perspective-taking, manipulated in a visuo-spatial sense, affects how well participants understand both positive and negative emotional facial expressions of others, and which consequences this has for the perceiver. In three pre-registered experiments (total N = 642), we tested whether perspective-taking affects the recognition (as assessed with a behavioral measure; i.e., the emotional multi-morph task) and perceived intensity (assessed via self-report) of a variety of facial expressions (disgust, sadness, joy, and surprise), as well as their emotional contagiousness (also self-reported). Compared to egocentric control trials, we found that visuo-spatial perspective-taking increased the perceived intensity of both positive and negative emotional facial expressions (Experiment 2), and the emotional contagiousness only of negative, but not positive, expressions (Experiment 3). Contrary to our predictions, visuo-spatial perspective-taking did not improve emotion recognition speed, and recognition accuracy was only slightly better after perspective-taking (Experiment 1). These findings challenge the notion that perspective-taking is an unambiguously beneficial social cognition by showing that while it intensifies the perception of all emotional expressions, perspective-taking causes emotional contagion only for negative emotional states. The potential of perspective-taking to improve and harm interpersonal relations is discussed in light of these findings.

ABSTRACT. Already in 1996, Thorpe showed that healthy individuals are able to categorize items even if the presentations were ultra-rapid (20ms). Furthermore, the enhanced perceptual hypothesis (Plaisted, 2000) assumes that this categorization processes may be impaired in individuals with autism spectrum disorder (ASD). This difficulty seems to be caused by abnormal categorization of atypical items, whereas semantic processing and the categorization process itself appear unimpaired. In accordance with the coarse-to-fine hypothesis, a typical member of a category can be recognized on a basic level, whereas an atypical item has to be processed on a subordinate level (Martinovic et al., 2008). Carmo et al. (2015) studied the process of categorization with very short presentation times (13 ms) and showed a different pattern of behavioral results for atypical items in longer presentations in individuals with ASD compared to healthy individuals, suggesting a malfunctioning feedback loop. To investigate this in more detail, in the current study we compared reaction times of two different categories (food/animals) with short (33 ms) and long (83 ms) presentation rates. Participants were asked to discriminate whether each item belongs to a specific category. We found differences in RT and d’ as a function of presentation times, item typicality, and item category. Notably, a specific RT advantage was observed for typical compared to atypical stimuli with long presentation times for individuals with ASDs (N = 15), but no such difference was evident in the matched controls (N = 20). Hence, these results support the notion of a malfunctioning feedback loop.

ABSTRACT. In the classic Stroop paradigm participants are asked to identify the color in which a color-word is written. Interference effects can be observed when the task-irrelevant color-word is mapped to a different response than the task-relevant color (inconsistent trials). Facilitation occurs when color-word and color are mapped to the same response (consistent trials). In contrast, in the Reverse-Stroop paradigm participants are instructed to identify the color-word, while ignoring the color. In this case, typically only interference effects can be observed. Obtaining both Stroop and Reverse-Stroop effects usually requires substantial changes to the paradigms. Here we present an experimental design which allows investigating both effects and their respective time courses with minimal changes in the procedure. Further, it is possible to manipulate the stimulus-onset asynchrony (SOA) between the task-irrelevant stimulus (word or color prime) and the target stimulus (color or word target). In a series of experiments (small-N design, 8 participants, 80 trials per cell and subject) Event History Analysis shows that (1) both Stroop and Reverse-Stroop effects can be observed, (2) Stroop and Reverse-Stroop tasks show similarly shaped reaction time distributions, (3) the effects increase with prolonged SOAs, (4) color information is processed faster than word information, (5) Stroop effects are time-locked to the onset of the task-irrelevant word stimulus, whereas (6) Reverse-Stroop effects are time-locked to the onset of the task-relevant word stimulus. Overall, we conclude that there is a speed advantage for the processing of color information, while word information has a stronger interference effect than color information.

ABSTRACT. According to the Metacontrol State Model, human behavior can be described in terms of two counteracting systems: one promoting persistence, the other promoting flexibility. The ability to shift the balance between these two opposing systems is referred to as metacontrol. Within recent years, an abundance of literature has demonstrated that metacontrol states are relatively flexible and can change as a function of both long-term (e.g., genetic profile, cultural background) and short-term factors (e.g., mood, meditation). The aim of the current project was to extend previous observations by investigating whether such states, once established, can be bound to environmental cues. In order to test this prediction, we designed four experiments (N > 55 for each study), each comprising an induction- and a test phase. Within the induction phase, participants were led to believe that stimulus processing, as measured in terms of performance on conflict tasks (e.g., Simon task), was modulated by the presence of contextual information. Herafter, participants were presented with a test phase, which was designed to test whether the mere presentation of previously shown cues allowed for a re-activation of the associated metacontrol state. Three out of four experiments provided converging evidence, showing that metacontrol states can be bound to contextual information. More specifically, the results indicated that the level of experienced conflict in the test phase was modulated by the associations formed within the induction phase. These results highlight the importance of contextual cues in shifting the persistence-flexibility balance towards one or the other cognitive control dimension.

ABSTRACT. A typical finding in response-interference tasks is the sequential congruency effect (SCE) characterized by a smaller congruency effect after previous incongruent trials compared to previous congruent trials. Traditionally, this effect has been accounted to top-down regulated cognitive control adjustments. However, this view has been challenged by theories of stimulus-induced retrieval of stimulus and response features. Here we tested an integrative account in which these theories play complementary rather than alternative roles: Abstract cognitive control states can be seen as a trial specific feature and thus are stored in trial specific event files. Therefore, control states can be retrieved in a subsequent trial when features overlap. Following this account, even repeating a response and conflict irrelevant context feature should retrieve the control state of the previous trial, leading to a larger SCE in context repeating trials as compared to context alternating trials. As predicted, in Experiment 1 (N = 39), a stronger SCE occurred when irrelevant context cues were repeated in subsequent trials of a prime-probe paradigm compared to alternating context cues. Importantly, the design eliminated stimulus - response repetitions and at the same time disentangled conflict and retrieval effects. The findings were conceptually replicated in Experiment 2 (N = 48) using a different stimulus set. The results provide strong evidence for a storage of control states in trial specific event files, together with stimulus, response, and context codes and thus promote an integration of top-down cognitive control and bottom–up episodic retrieval accounts of sequential conflict modulation.

ABSTRACT. I will outline a basic function of deliberation: that of decreasing the likelihood of expressing dominant responses. Furthermore, I will describe a minimalistic mechanism of how this function of deliberation could be implemented. To demonstrate the value of this function, I will reinterpret prior research in the area of decision-making and self-regulation that can be more parsimoniously explained by this function of deliberation. This reinterpretation includes attributing different decision-quality outcomes following deliberation to characteristics of the task instead of characteristics of deliberation. Finally, empirically, I will present three studies (N = 173) that use gambling choices with the specific task characteristic that dominant responses (operationalized by the size of framing effects) and decision quality (operationalized by expected value) are independent from each other. In line with the dominant-response reducing function of deliberation, inducing deliberation (as compared to spontaneity) reduced the likelihood of expressing dominant response while having no comparable effect on decision quality. The research will be discussed with respect to mechanisms of self-regulation and methodological issues regarding the attribution of observed behavioral effects following deliberation to characteristics of the deliberation process as compared to characteristics of the task context.

ABSTRACT. In an optimal stopping problem, people encounter a sequence of options and are tasked with choosing the best one; once an option is rejected, it is no longer available. The optimal solution is to choose the first number that is above a position dependent threshold. Recent studies suggest that people’s behavior in sequential search tasks is best described by a threshold strategy but their thresholds deviate from the optimal policy. We suggest that humans adapt their thresholds linearly, motivated by findings that humans tend to use linear functions to approximate more complex functions. We tested this hypothesis in an online shopping task and compared people’s strategy between three models that define how thresholds are generated. We show that linear thresholds provide the best account of the data. To replicate our results and to test whether we can predict how participants adapt to different environments we conducted a second experiment, where we manipulated the distributions of ticket prices across three conditions (scarce: left skewed, normal, plentiful: right skewed). The linear threshold model captures participants’ data accurately in all three conditions whereby the slope is adjusted to the environmental structure of the task. As predicted by the linear model, participants search longer in the plentiful than in the scarce environment. Overall, our work provides evidence that humans use linearity as a mental shortcut in optimal stopping tasks – understanding this heuristic enables predicting which environmental structures facilitate or impair human performance, providing a step towards a more complete theory of optimal stopping.

ABSTRACT. Our actions produce changes (i.e., effects) in the environment. The ability to detect these changes, to monitor these distal effects, is crucial not only in learning the relationship between actions and effects, but it is also a critical component of action control in the ideomotor framework. Thereby, effect monitoring is closely related to error monitoring, which also serves to monitor our own behavior and check for irregularities. In two experiments (n=48 each), we compared effect monitoring to error monitoring using a dual-task setup. Task 1 consisted of a three-choice button press, and action effects were displayed based on this response. Crucially, in some of the trials, a wrong effect was displayed after a correct response, and in some of the error trials, the correct rather than the erroneous effect was shown. Thereby, response-errors and effect-errors could be assessed separately. Task 2 was a simple discrimination task and served to measure the slowdown after response-errors and effect-errors. We found delayed responses after both, response-errors and effect-errors. Both these influences were additive, suggesting independent monitoring processes, one for proximal events, catching response errors, and one for distal events, checking for irregularities in the environment.

ABSTRACT. Spatial neglect is a debilitating neurological disorder marked by reduced attention to contralesional stimuli. We developed a gaze-contingent intervention in which eye movements to one visual hemifield were reduced over the course of several hundreds of trials by removing the stimuli in this hemifield whenever participants made eye movements towards it. The stimuli in the other hemifield were unaffected by eye movements. The aim of this study was to determine whether this intervention had an effect on the deployment of attention of healthy participants as a first step towards application in patients. Changes in attentional allocation were measured in a change blindness task. Having found strong effects in a first experiment (N=24) in both eye movement and reaction time data, a further aim was to establish the longevity of the effects. We compared two groups (N=25 each): the first group received the intervention once, the second group repeatedly on three consecutive days. The change blindness task was administered before the intervention and at four points in time after the last intervention (directly afterwards, + 1 hour, + 1 day, and + 4 days). The results show long-lasting effects of the intervention, most pronounced in the second group. Here the intervention changed the bias in the visual exploration pattern significantly until the last follow-up. We conclude that the intervention shows promise for the successful application in neglect patients.

ABSTRACT. While plenty of research suggests that numbers are represented spatially, some recent studies extended this to mental arithmetic by showing that addition/subtraction problems shift attention to the right/left. Two experiments investigate a) how these attentional shifts during the calculation phase develop over time by manipulating the delay between the arithmetic problem presentation and the spatial attention measurement (Exp 1 & 2) and b) how arithmetic task difficulty modulates these effects by varying the carry/noncarry-property of the arithmetic problems (Exp 2). In both experiments, spatial attention was measured via a temporal order judgment task (TOJ) where participants needed to decide which of two lateralized stimuli was presented first. The baseline consisted solely of the TOJ task and in the arithmetic task participants were first presented with the two-digit arithmetic problem via headphones and performed the TOJ task after the delay (250, 750 & 1500 ms) before responding to the arithmetic task. Operation and delay were varied within subjects (Exp 1 & 2) and the carry/noncarry property was varied between subjects (Exp 2). We found attentional shifts to the left/right for subtraction/addition problems compared to the baseline suggesting that visuospatial attention mechanisms are recruited during mental (symbolic) calculation. While Exp 1 showed no significant interaction between operation and delay, Exp 2 hinted at a decrease of the attentional shifts over time, but mainly for noncarry problems. We assume that two-digit carry-problems involve additional processes that either mask or don’t allow for an operation on the spatial magnitude representation that induces attentional shifts.

ABSTRACT. One puzzling result in research on value-driven attentional capture is that in learning-transfer paradigms, effects of value-associated stimuli on attention are often seen in transfer, but not in learning. In many of these studies, search tasks have been used in the learning phase. Here, we set out to cast some light on why there have rarely been reports of value-related effects on attention in the learning phase, but – somewhat magically – respective distraction effects then appeared in the transfer phase. Our hypothesis was that, during learning, two processes producing opposite effects cancel each other out: a speeding-up effect due to faster search, and a slowing effect due to more careful responding for stimuli associated with (higher) value. In four experiments with n = 24 each (N = 96), using search tasks and tasks without search, we indeed found evidence for both of these mechanisms: A value-dependent acceleration of target search, and a value-dependent slowing of the response to the target, which, however, dissipates over time. Together, these results show that the absence of value-associated effects during acquisition of the stimulus-value associations in the learning phase of two-phase learning-transfer paradigms is due to two antagonist mechanisms: Firstly, an increase in efficiency and, consequently, search speed for value-associated targets that pop out the more strongly the higher their associated value, and, secondly, an increased caution in responding with increasing value associated to the target in order to avoid losses.

ABSTRACT. Virtual reality (VR) might increase ecological validity as it allows to submerge into real-life experiences under controlled laboratory conditions. Visual attention research might particularly benefit from these methodological advances as previous studies have shown that cognitive demands differ significantly between 2D and 3D perceptual processes. This raises the questions if results gained by conventional 2D paradigms can be applied to real-world cognition. We studied sustained inattentional blindness by means of the invisible gorilla paradigm. Participants were either confronted with a 2D or a 3D360° VR video of two teams passing basketballs. In the classical 2D study, only ~30% of the participants noticed a gorilla entering the scene as opposed to the realistic VR-condition, in which the detection rate was increased to ~70%. Our results might be explained by a difference in perceptual load in the 2D versus VR-condition. In the 2D-condition, participants have to calculate depth- information, which conversely can directly be inferred from the binocular disparity in the realistic VR-condition, leaving resources for further attentional processing. The concept of sustained inattentional blindness thus might not be fully applicable to real-world conditions, challenging the standard assumption in psychological research that results from the laboratory generalize to real-world cognition.

ABSTRACT. When repeatedly exposed to simultaneously-presented stimuli, associations between the stimuli are nearly always established, within as well as between sensory modalities. Such associations have been shown to guide our subsequent actions and might also play a role in multisensory selection. The aim of the present study was to investigate the processing level of crossmodal distractor interference that is induced by association learning in multisensory selection. We assessed whether perceptual associations or rather response associations in a multisensory interference task are crucial for subsequent interference in a crossmodal interference task. Experiment 1 proved existence of a crossmodal aftereffect of multisensory selection. Experiments 2 and 3 then revealed that this effect depended on the perceptual associations between features rather than on the association between a feature and its response. Establishing response compatibility between multisensory and crossmodal task only did not lead to an interference effect (Experiment 2). On the other hand, feature associations without response compatibility (this was obtained by changing response effectors) were able to elicit subsequent crossmodal interference (Experiment 3). This pattern of results suggests that associations in multisensory selection and crossmodal distractor interference predominantly work at the perceptual, rather than motor, level.

ABSTRACT. Processing of visual stimuli varies with attentional task demands (i.e., requirements of stimulus selection). This can be seen in effects of the attentional set of a context task on response performance in another (probe) task, presented occasionally. Presenting the probe task after (invalid) cues indicating the upcoming context task Wendt, Kähler, Luna-Rodriguez, & Jacobsen, (2017, Front Psychol) obtained evidence for cue-based adoption of task-specific attentional sets (i.e., narrower focusing of visual attention after cues indicating an Eriksen flanker task than a task that required a homogeneous/heterogeneous judgment concerning a letter string). By contrast, the probe task performance did not yield evidence for persistence associated with the attentional set of task executed on the preceding trial. Here, we report two novel experiments, extending our previous findings to a novel type of stimuli and pursuing the lack of the persistence effect in more depth. In Experiment 1, we replicated both the cuing effect and the lack of a persistence effect concerning the deployment of attention to the global and local levels of hierarchical (Navon) stimuli. To investigate whether persistence of the attentional set was prevented by preparation for the following task (i.e., by cue-based replacement of the attentional set) we presented, in Experiment 2, probe task trials without cues. In this experiment, the overall probe task performance was affected in accord with the attentional demands of the predecessor task (visual-spatial attention in Experiment 2a and local-global processing in Experiment 2b). Together, our results indicate flexible, cue-based replacement of otherwise persisting attentional sets.