ABSTRACT. Two eye-tracking experiments are reported that examined the recognition of two-constituent compound words in alphabetic Finnish and logographic Chinese. In both languages, the majority of dictionary entries are compound words comprising two (or more) morphemes. In Finnish, two-constituent compound words vary greatly in length (aikaero = time difference vs. autovuokraamo = car rental), whereas in Chinese they are identical in length (时差 = time difference vs. 车行 = car rental). According to the visual acuity principle (Bertram & Hyönä, 2003), short Finnish compound words and all two-character Chinese compound words fit in the foveal vision and are thus recognized holistically, whereas long Finnish compound words are recognized via their components. The results are in line with the visual acuity principle. In Finnish, the effect of first constituent frequency indexing processing via components was significant for long compounds in gaze duration and probability of refixation, but not for short compounds. In Chinese, the first-constituent frequency effect was non-significant in gaze duration for the whole word and for the 1st and 2nd character. The Chinese results are also compatible with the Chinese Reading Model (Li & Pollatsek, 2020), according to which whole-word representations overrule the activation of individual characters.

ABSTRACT. Although known for reading, Sandy Pollatsek also explored how the effects of scene semantics affected attention, memory, and object recognition. From his early studies to much later ones, he used his knowledge of reading and eye movements to explore how parallel semantic associations may help to explain processing of scene images. Here, I will speak about his past work, the studies we conducted together on this topic and more recent work that was directly inspired from interactions with Sandy over the years. By examining how different aspects of scenes are processed over time, from the spatial layout to the identification of individual objects, I will show the over overlooked legacy of Sandy Pollatsek to the understanding of scene perception.

*Talk is part of the Symposium to honor Alexander Pollatsek’s legacy to eye movement research

ABSTRACT. This talk offers a quick snapshot of the travel from sensory information to letters and words during reading, one of the many topics tackled by Sandy Pollatsek. In 1998-1999, Sandy and I showed that one-letter different neighbors (space for the word spice) of a target word affected sentence reading. These insights yielded an ampler definition of a “neighbor” (e.g., addition or deletion-letter neighbors [horse vs. house], transposed-letter neighbors [calm vs. clam]). Related to the previous point and following Sandy’s ideas, it is critical to focus not only on English but also on very diverse orthographies. In this line, new efforts are devoted to clarifying the models' front-end by examining the role of diacritical vowels on word recognition and reading. We have recently shown that the omission of diacritics does not hinder (or only minimally) sentence reading in Spanish—in Spanish, diacritics only mark the stressed vowels. Critically, this pattern may change for those languages in which diacritics have other functions (e.g., vowel quality, length, tone, etc.). Further work in different languages and paradigms is essential to shed more light on the front-end of models of reading.

ABSTRACT. Datasets from eye tracking research during reading inevitably involves processing raw data into a form suitable for statistical analyses. For a reading experiment, this will typically include deciding whether to merge small fixations located close to each other, and whether to remove very short and/or very long fixations. Outlier analyses can include removing fixations 3 or 2.5 standard deviations from the grand mean, the overall participant mean or the mean of a participant for a specific condition. Subsequently, Linear Mixed Models are often reported on untransformed, log-transformed or using GLMMs which do not assume an underlying normal distribution (Lo & Andrews, 2015). Here we conducted a multiverse analysis, which consisted of running thousands of LMMs representing all reasonable combinations of choices in cleaning and analysing fixation data. In this talk, we will explore the impact of these choices focusing on the reported size of the frequency effect during a reading experiment using an EyeLink 1000. Amongst our findings is that the frequency effect in single fixation durations can vary by more than 5 milliseconds depending on the data processing choices made, even though we restricted ourselves to cleaning and analysis methods that would be considered acceptable practice by researchers.

ABSTRACT. ****ABSTRACT SUBMITED AS PART OF THE SYMPOSIUM IN HONOUR OF ALEXANDER POLLATSEK****

To mark the immense contribution to the field of eye movements and human visual cognition that Sandy Pollatsek made during his career, I will try to deliver a talk that has an empirical basis, challenges current thinking, and hopefully, stimulates discussion (all things he valued). I will consider how visual and cognitive processes are operationalised over linguistic units across fixations during natural reading. The empirical basis of the talk will be results from a series of eye movement experiments that my Chinese colleagues and I have recently conducted to explore the Multi-Constituent Unit Hypothesis (Zang, 2020). These experiments have caused us to think that processing occurs serially and sequentially, but with respect to lexical units that may be larger than individual words. Such a suggestion might apply for unspaced languages (e.g., Chinese), agglutinate languages (e.g., Finnish) as well as (non-agglutinate) spaced languages (e.g., English). Additionally, I will suggest that current models of lexical processing do not adequately map onto natural reading circumstances. At a minimum, ecologically valid models of lexical identification require processing mechanisms to support successive (fixation-by-fixation) episodic delivery of visual information that is modulated by acuity limitations based on (launch and landing) fixation sites.

ABSTRACT. Our theoretical understanding of eye glance behaviors in scene recognition, visual search, and attention has advanced considerably since the 1970s. Our understanding of the role of eye glance behaviors in the more practical tasks such as reading has also advanced. Much more recently, those behaviors have been shown to be critical to understanding just how engaged drivers are in the dynamic driving task, why novice drivers are more likely to crash than experienced drivers, why drivers with ADHD are more at risk than neurotypical drivers, why older drivers are especially at risk in intersection crashes, and whether drivers can be trained to decrease their crash risk, to name just a few examples. No one has put together the various theoretical advances with the more applied advances. Alexander Pollatsek has been critical to this integration. His many seminal contributions to this integration are discussed and highlighted.

ABSTRACT. As a logographic writing system, Chinese has many unique properties compared with alphabetic writing systems such as English. One important unique property is that there are no interword spaces to demarcate words in Chinese. This raises important questions regarding how Chinese readers segment words and how they choose their saccade targets. To understand how Chinese readers conquer this challenge, Sandy Pollatsek and I conducted a series of experiments together in the last 15 years. Based on the results of these findings, we also constructed a computational model of word processing and eye movement control during Chinese reading (CRM). These works have been valuable to understand the unique mechanism of Chinese reading. In this talk, I will share some of my memories on working with Sandy on these projects.

ABSTRACT. Can you efficiently look for something even without knowing what it looks like? According to theories of visual search, the answer is no: A template of the search target must be maintained in an active state to guide search for potential locations of the target. Here, we tested the need for an active template by assessing a case in which this template is improbable: the search for a familiar face among unfamiliar ones when the identity of the target face is unknown. Since people are familiar with hundreds of faces, an active guiding template seems unlikely in this case. Nevertheless, participants were able to guide their gaze towards the target as long as extrafoveal processing of the target features was possible. Additionally, individual's differences in the ability to process familiarity through extrafoveal vision were correlated with proficiency in the search task. These results challenge current theories of visual search by showing that guidance can rely on long-term memory and extrafoveal processing rather than on an active search template.

ABSTRACT. Have you recently failed to notice your roommates’ new poster or plants, or worse – your friends’ new haircut? We hypothesized that such failures are the result of attenuation of visual exploration that happen due to an internally stored memory that is the subject of our investigation. In order to test this hypothesis, we designed an experiment including either visually familiar or non-familiar stimuli, in situations of encoding and retrieval. We show that both during encoding and retrieval, a unique pattern of gaze behavior towards the familiar percept is reflected by fewer and longer fixations. Thus, the results suggest that visual exploration is indeed attenuated as the informational gaps are filled by accessing existing mental representations. This research sheds light on the relationships between internal and external exploration mechanisms by bringing forward the tradeoff between the two processes.

ABSTRACT. Watching cinema involves interpreting characters’ mental states through a process known as Theory of Mind (ToM). To investigate these processes and how the viewer’s mental model of the character’s mental state influence how they attend to and process the film, we focused on dramatic irony scenes, where the audience knows something that one of the characters does not. In a between-subjects design, we provided critical information that the characters did not possess to some viewers (dramatic irony group) and not to others (control group, which had the same information as characters), thereby creating different event models of the characters’ mental states. We hypothesized that if the dramatic irony group build more complex ToM event models that these processing differences will be evident during the film viewing as deeper cognitive processing. Combining eye tracking with a self-paced film-viewing task we will examine: (a) pupil dilation as an indirect marker of processing load and (b) the analysis of spatio-temporal distribution of overt attention, i.e. what part of the images participants are processing differently across conditions, informing us about how participants extract information about characters’ mental states, and whether there are longer viewing times when event models are switched/updated to follow character perspectives.

ABSTRACT. Since after covid pandemic began, the population of patients with depression has been skyrocketing. Sometimes depression is hard for patients to recognize if they have it or not by themselves, especially in the early stage. Our main goal is to develop systems to detect the early stage of depression using eye movements data while they are talking to virtual avatars or humans such as their friends, families, or colleagues, in their daily life.

We will report the result which focused on whether eye movements reflect depression while people are talking to virtual avatars in the non-clinical interviews. We collected eye-tracking data of participants with or without depression while they were interacting with a virtual avatar or human interviewer and computed statistical analysis. As the result, eye movements are possible to be one of the criteria to classify people as being with or without depression.

 This research was supported by European Regional Development Fund (ERDF) for Post-doc projects grant agreement No 1.1.1.2/VIAA/4/20/668.

ABSTRACT. Reading is increasingly taking place on digital support, which is vector of interruptions that disrupt the attention during reading. This study reports the consequences of an interruption on reading behavior, and text comprehension for the information being read precisely when the interruption occurs. 38 participants read four long procedural texts while we recorded their eye-movements (mean number of fixations, regressive fixations and mean fixation duration). The readers were interrupted by an arithmetic verification task either in the middle of the paragraph (intra-paragraph condition) or between two paragraphs (inter-paragraph condition). In the intra-paragraph condition, the eye movement analysis revealed more re-reading behavior when an interruption occurred. We assessed comprehension with 48 questions and separated our participants into two groups (high and low comprehenders) based on the median of the error rate. The high comprehenders showed more re-reading behavior (fixation and regressive fixation) in the sentence including the target information than low comprehenders. However, independently of the position of the interruption (inter or intra-paragraph) the comprehension performances were not degraded. Those results are discussed in regard to long term working memory (Ericsson & Kintsch, 1995).

ABSTRACT. The concept of desired difficulty assumes that perceived increased task difficulty may lead to deeper processing and benefit memory performance. During reading, this idea can take the form of using typographically altered fonts, e.g., Sans Forgetica (SF), a typeface specifically designed for this purpose. The present study examines the impact of these letter alterations in terms of moment-to-moment processing and text comprehension. Participants read non-fiction texts while their eye movements were recorded. After practice with the new font, each participant read two texts in Arial and two in SF. Following the Construction-Integration-Model (Kintsch, 1988), text comprehension was calculated on the levels of propositional text base and situation model. We found a substantial inflation of early and late viewing time parameters when reading the SF font. These effects were most prominent at the beginning of the experiment. Interestingly, an expected interaction between font type and word length was lacking. In addition, no difference on either level of comprehension was found between the two fonts. Apparently, linguistic processing and oculomotor control was adjusted to meet a certain comprehension criterion. We show how this global adjustment (Radach, Huestegge & Reilly, 2008) is being implemented in local fixation patterns and discuss theoretical implications.

ABSTRACT. Doubly quantified sentences like “Every kid climbed a tree” are ambiguous regarding whether there is one tree or multiple trees. Previous research has shown that grammatical factors (e.g., the order in which the quantifiers appear) do not affect the general preference for a singular entity, evidenced in shorter reading times for “The tree was” than “The trees were” in a subsequent sentence. However, the influence of contextual information has received little attention. In the current study, 48 younger (18-30) and 48 older adults (65+) had their eye movements monitored while they read sentences in which the context was biased towards either singular (e.g. “Every student in the class is listed on a register.”) or plural entities (e.g. “Every school pupil in the country has their attendance marked on a register.”). A singular (“This register…”) or plural continuation (“These registers…”) then followed. Results showed that contextual bias influenced on-line linguistic processing, with longer reading times for continuations which were incongruent with the context. Also, evidence of greater processing difficulty was identified when contextual factors were in opposition with grammatical factors, especially for older adult readers. The present findings support a parallel constraint-satisfaction approach under which multiple factors interact during on-line processing.

ABSTRACT. While many studies have examined the mechanisms underlying reading for comprehension, relatively few have directly examined the effects of readers’ goals on eye movement behaviour (e.g., Fitzsimmons et.al., 2020; Strukelj & Niehorster, 2018). Here we present three experiments that specifically examined how reading goals modulate the processes involved in sentence integration. Specifically, we investigated how different goals (reading for comprehension versus scanning for a topic [Experiment 1 & 2], skimming for gist versus scanning for a topic [Experiment 2]), modulate effects of syntactic ambiguity (Experiment 1), semantic plausibility (Experiment 2) and sentence wrap-up (Experiment 3). Reading times were longer for reading compared to scanning for a topic, and longer when skimming for gist compared to when scanning for a topic. Crucially, the results showed that effects of ambiguity, plausibility and sentence wrap-up were similar during the initial processing of critical words across all reading goals. However, for measures sensitive to rereading, (regression path duration), effects were larger for reading compared with scanning for a topic. The results build on previous work (Weiss et. al., 2018) suggesting that sentence integration processes can be modulated by task demands. We discuss the implications for theories of eye movement control during reading.

ABSTRACT. The parallel processing OB1 Reader Model (Snell et al., 2018) postulates that multiple words are lexically processed in parallel and can be assigned to their correct position by a cognitive mechanism to form a valid sentence frame regardless of their order of presentation during natural reading. In this talk, we present an eye movement study that tested how word transpositions and grammatical well-formedness affect reading. In the current experiment, we presented 5-word sentences in which the positions of the third and fourth words (the target pair of words) and the identity of the final word (the post-target) were orthogonally manipulated to produce grammatical violations. Our results showed that the word transposition did not cause any disruption to the processing of the first two words in the sentence, consistent with the suggestion that lexical and syntactic information was not extracted from the parafovea and that words are lexically processed in a serial rather than parallel fashion. Moreover, the transposition produced significant disruption on the third word of the sentence, with longer reading times when the two targets were transposed compared to not transposed, suggesting that the order in which words are presented is important for natural reading.

Snell, J., van Leipsig, S., Grainger, J., & Meeter, M. (2018). OB1-reader: A model of word recognition and eye movements in text reading. Psychological Review, 125(6), 969- 984. https://doi.apa.org/fulltext/2018-37844-001.html

ABSTRACT. Parafoveal-on-foveal (PoF) effects, in which a parafoveal word (n+1) influences processing of the foveal word (n), have previously been interpreted as possible support for parallel instead of serial processing during word recognition. Here, we document the influence of an emoji’s meaning on its preceding target. We examined eye movements while participants read sentences containing a target word (e.g., coffee in the sentence “I enjoyed my tall coffee”) that was immediately followed either by no emoji, a semantically congruent (e.g., coffee cup) or an incongruent (e.g., beer mug) emoji. First-pass fixation durations were shorter on the foveal target word (n) when the parafoveal emoji (n+1) was semantically congruent rather than incongruent (i.e, an emoji-elicited semantic PoF effect), which suggests that emojis and text can potentially be processed in parallel. Building on these results, we conducted E-Z Reader simulations to explore if a serial model of eye movement control could be extended to emojified text. Our simulations indicated that emojis require more time for perceptual and cognitive processing compared to three-letter nouns that were equated to the emojis in length. We discuss how models of eye movement control during reading could be modified in the future to accommodate emojified text.

ABSTRACT. Restoring vision after congenital visual deprivation often results in incomplete functional recovery. Here we tested whether early visual experience is necessary for the development of systematic visual exploration. A free-viewing paradigm using real-world images was used to investigate the exploratory behavior of ten individuals born with bilateral dense cataracts whose sight had been restored after a prolonged period of severe visual deprivation. Their behavior was compared to three control groups: normally sighted age-matched controls, developmental cataract reversal individuals, and individuals with infantile nystagmus without a history of visual deprivation. Sight-recovery individuals presented moderate-to-severe visual impairments and gaze instability (nystagmus). However, their visual exploration patterns were successfully predicted by those of normally sighted controls and were indistinguishable from individuals with infantile nystagmus. Similar to all control groups, visual exploration in sight-recovery individuals was based on the low-level (luminance contrast) and high-level (object components) visual content of the images. Sight-recovery individuals’ systematic visual exploration was associated with better object recognition, suggesting that active vision might be a driving force for visual system development and recovery. Overall, the present results argue against a sensitive period for the development of the neural mechanisms associated with active visual exploration.

ABSTRACT. Introduction

People with homonymous hemianopia (HH) can use scanning strategies that compensate for their visual field defect to overcome difficulties with mobility activities such as walking. However, environmental demands may influence the use of compensatory scanning. This study aims to examine the effect of increased environmental demands on compensatory scanning in people with HH.

Method

Ten participants with HH and ten with normal vision walked for 2 minutes in a quiet neighbourhood and a shopping centre while their gaze was recorded using a mobile eye tracker (Pupil Invisible). For both participant groups, we evaluated scanning behaviour in terms of exploration, length of scans, and distribution of scanning.

Results

Our preliminary results show that both people with HH and with normal vision increased their distribution of scanning and exploration when walking in the shopping centre compared to the quiet neighbourhood.

Conclusion

During walking, people with HH as well as people with normal vision adapt their compensatory scanning depending on environmental demand. Future research should investigate whether and how people with HH benefit from these changes in compensatory scanning and examine differences in changes in scanning behaviour between people with HH and normal vision.

ABSTRACT. Previously, transgender people have been found to experience more body dissatisfaction in comparison to cisgender people. A link between attention, body perception and body dissatisfaction has been found but the role of eye movements is unclear. Particularly, it is not known if attention is deployed to particular body parts. Therefore, this study aimed to determine whether transgender and cisgender people show an attentional bias toward body parts they are dissatisfied with. To test this aim, we recruited 90 participants (30 transgender, 30 cisgender males and 30 cisgender females) and tracked their eye movements while viewing 3D scans of bodies. Participants were unaware of the true aim of the experiment and were asked to perform a body recognition task (learn a set of 12 3D bodies from different viewpoints and then perform a 2AFC recognition task). After this task, they were asked to complete a self-report questionnaire about body (dis)satisfaction. A novel method to map eye movement onto 3D bodies was used, allowing us to aggregate data across bodies and viewpoints. The three groups of participants performed similarly during the body recognition task, but significant differences were found when comparing gaze patterns between groups. Most importantly, we found a significant direct correlation between body dissatisfaction and fixations patterns, across all groups, but particularly for the transgender and female group. We found that the more a participant disliked a body part (e.g. the stomach area) the more they fixated that body part showing a clear cognitive bias towards disliked body parts.

ABSTRACT. The motor learning process underlying the identification of a periodic stimulus and initiation of movements to synchronize to its arrival is known as temporal prediction. Temporal prediction is needed for motor coordination, yet little is known about its timeline of maturation across typical and atypical neurodevelopment. Here, we recorded saccades to regular and irregularly timed visual targets in typically developing youth (N=115; aged 6-24 years) and in adolescents with psychiatric diagnoses of the impulsivity domain (N=44; aged 11-18) to elucidate the typical development of predictive saccades (saccade reaction time [SRT] <90 ms; in anticipation of visual target) to 5 interstimulus intervals and compare behaviour among healthy and psychiatric groups. In addition to saccade metrics (e.g., reaction time, peak velocity, amplitude), we recorded participants’ pupil size and blinks throughout the tasks to query how these behaviours varied according to whether participants predicted or reacted (SRT > 90 ms) to periodic targets. Indeed, when predicting targets, saccades were hypometric, pupil size was reduced, and blinks were suppressed prior to target appearance. The age of maturation for predictive saccades was ~14. Psychiatric adolescents, compared to age-matched controls, produced an increased frequency of predictive saccades towards irregular targets.

ABSTRACT. Saccades may provide effective biomarkers for neurodegenerative disease due to overlap between oculomotor circuitry and disease-affected areas. Accurate characterization of disease-related saccade behaviour, and its links to global and specific aspects of cognitive impairment, is crucial to biomarker development. To achieve this, we used data from the Ontario Neurodegenerative Disease Research Initiative, which recruited multiple large, directly comparable cohorts of patients with Alzheimer’s disease, mild cognitive impairment, amyotrophic lateral sclerosis, frontotemporal dementia, Parkinson’s disease, or cerebrovascular disease. Patients (n=450, age 40-87) and healthy controls (n=149, age 42-87) completed a randomly interleaved pro- and anti-saccade task; patients also completed an extensive neuropsychology battery. We explored the relationships of saccade parameters (e.g. task errors, reaction times) to one another and to neuropsychology-based cognitive domain scores (e.g. executive function, memory). Task performance consistently worsened with global cognitive impairment; subsets of saccade parameters were interrelated and also differentially related to cognitive domain scores (e.g. antisaccade errors and reaction time associated with each other and with executive function). This suggests the subsets may index temporally distinct brain processes that connect saccades to cognition and neurodegeneration, and may have implications for use of IPAST as a cognitive screening tool in these diseases.

ABSTRACT. Restoring functional sight requires that the electrical stimulation should convey information to the brain that is associated with the correct spatial location in the scene. A visual cortex stimulator can bypass the eye and the optic nerve and create phosphenes (perception of light) without light entering the eyes. Nonetheless, recently we demonstrated that eye movements dominate the perceived location of cortical stimulation-evoked phosphenes, even after years of blindness. In the current presentation we will present results from patients demonstrating the correlations between eye movements and the visual percepts. We instruct patients to conduct an eye movement toward the phosphene and to use eye movements as a marker to construct the spatial map of the implanted electrodes. Experiments were conducted with blind patients implanted with the NeuroPace Responsive Neurostimulator (RNS) and the Orion visual cortical prosthesis devices. In contrast to a retinal prosthesis, in a cortical visual prosthesis, the layout of the implanted array does not match a retinotopic map and it is necessary to find the visual-field location of the percept of each implanted electrode. To establish the spatial map of the electrodes, users were instructed to conduct an eye movement to the location of the phosphene generated by electrical stimulation of the occipital lobe. Two different schemes were compared. In the first, a brief stimulation was presented and the subject moved their eyes after the end of the stimulation toward the phosphene’s remembered location. In the second, a longer stimulation was presented and the subject moved their eye during the stimulation to track the phosphene’s location. In the latter case, because the stimulation is continuously mapped based on eye position, an eye movement during the stimulation caused the phosphene to move. Results show that subjects were able to conduct a smooth pursuit motion as a result of constant stimulation. These experimental setups demonstrate that the integration of eye-tracking recording can be used to create the spatial map of a cortical visual implant.

ABSTRACT. Language learners often struggle reading in the target language due to a lack of vocabulary (Grabe & Stoller, 2013). Effective vocabulary coping strategies are essential and include three main options: trying to infer word meaning, ignoring the word, and dictionary use. However, educators have differing philosophies on the most effective strategies, and learners may be overly reliant on a certain strategy. Prior research on this area has been dominated by surveys and think-aloud protocols. However, participants may not be fully aware of their strategy use, or they may report based on what they perceive is the ideal response.

This presentation overviews how eye tracking has been utilized to more empirically examine vocabulary coping strategies. Two studies involved Japanese college-aged learners of English (N = 57 and 39) in task-based reading. Eye tracking revealed that, in cases where a novel word was mid-sentence, most participants did not even fixate on context cues following the word before checking the dictionary. Participants who made regressions after fixating on a pseudoword to re-read the context before looking up the word performed better on the post-reading tasks. Other eye movement data and implications for researchers and educators will be overviewed.

ABSTRACT. In the majority of arguments used in Chilean primary school textbooks, major premises -the ones that guarantee the logical link between the minor premises and the conclusion- are kept implicit. This pattern slows down processing and hinders comprehension, especially among poor readers (Malaia, Tommerdahl & McKee, 2015; Prado, Spotorno, Koun, Hewitt, Van der Henst, Sperber, Noveck, 2015). To explain the phenomenon, it has been hypothesized that, faced with the absence of the major premise, readers do not elaborate the logical connections (Britt, Kurby, Dandotkar & Wolfe, 2008). Given that readers may have an insufficiently precise representation of the text, discourse markers may facilitate this reading (van Silfhout, Evers-Vermeul & Sanders, 2015). To assess whether this is the case, Chilean 6th and 8th graders read short argumentative texts while their eye movements were recorded. Each text included one of four conditions, considering two levels for each variable: major premise (implicit and explicit) and marker (implicit and explicit). Preliminary results indicate that the target sentences, minor premise and conclusion, were read faster and adequately understood with the conditions explicit premise-connector and implicit premise-explicit connector. This means that there was no effect on the degree of explicitness of the major premises. The explicitness of this premise does not result in more efficient processing of arguments, nor does it guarantee adequate comprehension, as compared to the presence of connectors.

ABSTRACT. Skilled readers typically identify words more accurately when fixating them slightly left of the central character, the so-called optimal viewing position. There are two main explanations for this effect, which are not mutually exclusive. The first claims that the optimal viewing position lies left-of-center due to the particular constraints of the human perceptual system. The second explains the effect in terms of how words compete with each other; specifically, the beginnings of words tend to be more unique and therefore more informative about word identity, making a left-of-center fixation more advantageous. We explore this effect through the lens of a Bayesian cognitive model and two experiments using artificial lexicons in which we can carefully control how information is distributed across wordforms. Our results suggest that readers are sensitive to the distribution of information, targeting different positions depending on whether the language they learned is more informative on the left or right. Furthermore, readers do not simply target the position that contains the most information; rather, they target the position that will yield the best view of the word overall, accounting for both information distribution and the asymmetry of the human visual span.

ABSTRACT. Binocular coordination (saccade disconjugacy, vergence drifts and fixation disparity) and binocular advantages in reading (shorter total reading times and fixation durations in binocular compared to monocular reading) typically vary with the amount of individual, horizontal heterophoria, i.e., a resting position of vergence eye movements. Since prismatic glasses (identified via MCH-method) change heterophoria, we investigated whether prismatic glasses also affect different aspects of binocular coordination during reading. We collected binocular eye movement data (Eyelink II) for 54 participants in four reading conditions (30 sentences each): 1) monocular, 2) binocular, 3) binocular with added disparity (caused by MCH prisms) and 4) binocular with added disparity after 6 months. During the 6 months delay, 29 participants wore refractive glasses including prismatic effects (mean heterophoria: 2.1 pdpt) and 25 participants wore refractive glasses only (mean heterophoria: 2.5 pdpt). While all participants showed some binocular advantage for all binocular reading conditions, the binocular advantage tentatively increased further for our prism group after 6 months. Furthermore, saccade disconjugacy increased slightly and objective fixation disparity was significantly reduced after 6 months. Thus, after some time, individual prismatic glasses (MCH) changed aspects of the binocular coordination in reading and slightly increased the binocular advantage for our heterophoric participants.

ABSTRACT. While there is growing evidence that reading comics is sequential and ordered, like reading, we currently lack an understanding of the processing mechanisms that underlie comic reading. A key issue is the influence of peripheral information on the reading process. Using the moving-window and invisible boundary paradigms, across three studies we investigate whether comic reading is disrupted when peripheral information is removed in single line comic strips (Experiment 1) and multi-line comics (Experiment 2); as well as what more detailed image information is extracted from upcoming panels (Experiment 3). The studies show that readers are aware of the content of at least the next panel, extracting both morphological and semantic information from the images, and identifying the location of text before moving to the next panel. The findings add further evidence to the idea of reading comics as an ordered process, and the existence of a visual language, with its own narrative grammar.

ABSTRACT. Observers react faster when the target appears on the same object as a precue than when they appear on different objects. The ‘Attentional shifting’ account of such object-based effects underscores the higher cost of shifting attention between objects than within an object. The ‘Attentional spreading’ account attributes these effects to automatic spreading of attention along the object. We tested these accounts using attentional modulations of the pupillary light response (PLR). To test the spreading account, we employed objects composed of white-to-gray and black-to-gray luminance gradients. If attention spreads along the objects, pupil size should differ when the different types of objects are cued. To test the shifting account, we displayed a single object that could be black or white. A precue appeared inside or outside the object, and a target followed at the same or different location. If disengaging attention from a location inside the object is slower, then changes in PLR should start later when the precue appears inside the object and the target appears outside the object. We found evidence for the attentional shifting account but not for involuntary spreading of attention. Evidence for attentional spreading along the object emerged only when such spreading matched the observers’ goals.

ABSTRACT. The amplitude of the phasic pupil dilation response (PDR) is assumed to be independent of luminance-related baseline pupil diameter. There is, however, little empirical support for this assumption. Instead PDR amplitudes have been reported both to decrease and to increase with baseline diameter. Here, we systematically examined this relation in an auditory oddball paradigm including rare to-be-detected target and rare task-irrelevant novel sounds among frequent standard sounds. Display luminance was manipulated in six blocked conditions between 0.3 and 100cd/m2. The amplitude of the early parasympathetic inhibition-related principal component of the PDR decreased linearly with increasing luminance-related baseline diameter. The late sympathetic activation-related component showed an inverse u-shaped relationship with small PDR amplitudes at low and high luminance levels, but large amplitudes at medium luminance. For the spontaneous, non-luminance-related fluctuations of baseline pupil diameter, we observed decreasing amplitudes of both PDR components with increasing baseline. Importantly, this decrease was stronger for task-irrelevant novel sounds compared to task-relevant target sounds. We discuss possible explanations for the observed results in terms of the dynamics and mechanical properties of the pupillary system. Divisive baseline correction is discouraged in pupillometry as the underlying assumption does not hold. We propose regression-based baseline correction as feasible alternative.

ABSTRACT. We investigated system drift by tracking firmly mounted artificial eyes using five different eye trackers; SMI HiSpeed 240, SMI RED250Mobile, SR Research EyeLink 1000+, SR Research EyeLink II, and Tobii Pro Spectrum. The system drift in the gaze signal ranged from 0.03 to 2.71 degrees over two hours, depending on the eye tracker. A closer investigation of the eye images from the SMI HiSpeed 240 uncovered one potential mechanism for this, with images changing brightness over time. We show that this change in eye-camera images affects the contour estimation of the features in the eye image (pupil and corneal reflection), resulting in system drift not only for the gaze position but also for the pupil size. The artifactual changes in pupil size are on the same order as the largest cognitively driven pupil changes and could be a concern for experiments focused on temporal effects, such as learning, fatigue, or vigilance. The system drift was observed after starting the eye tracker from room temperature, but was considerably less noticeable after about an hour of use. Based on our findings, we recommend warming up eye trackers for at least one hour before recording data.