Eye Movements: Perceptual advantages and disadvantages

Talk Session: Tuesday, May 20, 2025, 5:30 – 7:15 pm, Talk Room 1

Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions

Talk 1, 5:30 pm

Exploring the limits to the “illusion of relative stability” of images slipping consistent to fixational eye motion

Josephine C. D'Angelo¹ (), Pavan Tiruveedhula¹, Raymond J. Weber², David W. Arathorn², Jorge Otero-Millan¹, Austin Roorda¹; ¹University of California, Berkeley, ²Montana State University

The human visual system is exquisitely sensitive to detecting relative motion; this ability requires the presence of world-fixed images in the scene which serve as frames of reference. However, images moving in the direction consistent with retinal slip appear to have little to no motion relative to world-fixed retinal image background content, suggesting that under this special condition perceptual stabilization overrides motion sensitivity (1). This phenomenon was called the “Illusion of relative stability.” We aimed to explore the limits to this illusion by measuring the perceived motion of images slipping opposite to and in the same direction as eye motion with magnitudes closer to the transition between directions. For example, would a stimulus moving in a direction consistent with retinal slip but only slipping with 10% of the retinal motion of a world-fixed stimulus also appear relatively stable? We used an adaptive optics scanning light ophthalmoscope to present stimuli that moved contingent to fixational eye motion. To quantify the perceived motion of these stimuli, we presented a second stimulus that moved in a random trajectory and the subjects adjusted its magnitude of motion until it appeared perceptually equivalent to the retina-contingent stimulus. We found that images slipping opposite to eye motion appear less in motion while images moving in the same direction appear to have a higher magnitude of motion which on average is similar to the image’s magnitude of world motion. Even stimuli slipping with only 10% of the retinal motion of a world-fixed stimulus appear relatively stable. These results suggest a discontinuity in motion perception which is dependent on the direction that images move contingent to fixational eye motion. (1) D'Angelo, Josephine C et al. “A paradoxical misperception of relative motion.” Proc. Natl. Acad. Sci. U.S.A. vol. 121,48 (2024): e2410755121. doi:10.1073/pnas.2410755121

NIH R01EY023591; NIH T32EY007043; NIH R00EY027846; Berkeley Center for Innovation in Vision and Optics

Talk 2, 5:45 pm

Temporal predictions underlie the extra-foveal preview effect across saccades

David Melcher¹, Michele Deodato¹; ¹New York University Abu Dhabi

In natural viewing, we look around objects in the scene. Thus, visual processing of objects typically involves a time course in which we make a saccadic eye movement to a target object that was first viewed, and selected, using extra-foveal vision. Previous studies have shown that this extra-foveal preview can strongly influence perceptual judgments about that stimulus after it is fixated. However, the mechanisms underlying trans-saccadic perception remains a matter of debate. Here, participants performed a gaze-contingent task in which an extrafoveal saccade target stimulus (the preview) was replaced with a tilted stimulus (the target) during a saccade directed to it. On separate trials, the preview stimulus was either identical to the target (valid preview) or different (invalid preview). We ran multiple experiments, with invalid previews consisting of faces with different identities, inverted versus upright faces, or faces versus houses. Critically, on some trials we added a brief, blank delay at the beginning of the post-saccadic fixation period, before the appearance of the target. The added temporal gap at the beginning of the new fixation dramatically modulated the preview effect, leading to an elimination or even a reversal of the usual preview validity benefit. Overall, these findings suggest that saccadic programming and processing of the extrafoveal saccade target creates a category-level and temporal prediction which underlies the parafoveal preview effects in trans-saccadic perception.

This work was supported by the NYUAD Center for Brain and Health, funded by Tamkeen under NYU Abu Dhabi Research Institute grant CG012. Part of the work was conducted at the Brain Imaging lab within the Core Technology Platforms at NYU Abu Dhabi

Talk 3, 6:00 pm

Voxel-wise predictive encoding models reveal evidence for pre-saccadic remapping in the human visual cortex

Yong Min Choi¹ (), Julie D. Golomb¹; ¹The Ohio State University

Retinal visual inputs shift drastically across saccadic eye movements. To align pre- and post-saccadic visual information and aid perceptual stability, neurons’ receptive fields (RFs) are predictively remapped in anticipation of upcoming saccades toward the future RF location (forward remapping) or the saccade target location (convergent remapping). Despite extensive evidence of pre-saccadic RF remapping in non-human primates, there remain open debates about the nature of remapping in different brain areas, and, more generally how remapping operates in the human visual system. In the current study, we developed a novel fMRI paradigm to investigate pre-saccadic RF remapping in the human brain using a hypothesis-driven, voxel-wise predictive encoding model approach. Adult participants completed two fMRI sessions: (1) a static population receptive field (pRF) mapping session to estimate voxel-wise static pRFs, and (2) a main experiment session presenting noise-like visual stimuli either during stable fixation or during saccade preparation. Using each voxel’s estimated static pRF, we constructed different hypothetical models of pre-saccadic remapping (e.g., “no predictive remapping”, “forward remapping”, “convergent remapping”) and evaluated how well each model predicted actual neural responses in the main experiment session. During stable fixation, as expected, the neural activity of most voxels in visual areas was best predicted by models assuming no predictive remapping. Critically, during saccade preparation, many voxels showed improved predictability for models incorporating forward and/or convergent RF remapping. These findings provide novel evidence for pre-saccadic remapping at a voxel-level across the human visual cortex, with this new approach offering exciting potential to broaden our understanding of how RF remapping operates across the brain and links to behavior.

NIH R01-EY025648 (JG), NSF 1848939 (JG)

Talk 4, 6:15 pm

Pre-saccadic attention in Parkinson’s Disease: effects of aging and dopamine level

Oliver L. Steiner^1,2,3 (), Nina M. Hanning^3,5, Sarah Melchert¹, Sven-Florian Jaeger¹, Fabian Klostermann^1,2, Martin Rolfs^2,3,4; ¹Klinik für Neurologie mit Experimenteller Neurologie, Charité – Universitätsmedizin Berlin, Deutschland, ²Berlin School of Mind and Brain, Humboldt-Universität zu Berlin, Deutschland, ³Institut für Psychologie, Humboldt-Universität zu Berlin, Deutschland, ⁴Bernstein Center for Computational Neuroscience, Berlin, Deutschland, ⁵Department of Psychology and Center for Neural Science, New York University, New York, NY 10012, USA

Navigating dynamic and unpredictable environments requires allocating attentional resources for impending movement. While mechanisms of pre-saccadic attention – the anticipatory enhancement of sensory processing at an imminent saccade target – have been characterized in cortical networks, the role of subcortical structures remains largely unexplored. Based on the important role of the basal ganglia for attention allocation and movement preparation, we hypothesized that dopamine depletion in this subcortical network disrupts the temporal dynamics of pre-saccadic attention. To test this, we assessed patients with Parkinson's Disease (PD) – a neurological movement disorder characterized by dopamine degeneration in the basal ganglia – both ON and OFF dopaminergic medication. We compared their performance to age-matched controls and young observers.

 Participants fixated on a central bull's-eye while two placeholders, each consisting of four black dots, were positioned 10° left and right of fixation. Upon stable fixation, a central line cue indicated the saccade target. Participants were instructed to "look as fast and precisely as possible" at the indicated placeholder. At varying times from approximately 200 ms before saccade execution up to saccade onset, we presented ±45 degree oriented 1/f noise for 83.3 ms, embedded in a dynamic stream of random 1/f noise. Once participants had executed the saccade, they verbally reported the stimulus orientation (clockwise or counterclockwise from vertical). 

Both elderly controls and PD patients exhibited prolonged saccadic latencies compared to young observers, accompanied by a flatter slope of attentional enhancement approaching saccade onset. PD patients showed a reduced pre-saccadic attentional benefit immediately before saccade execution. No significant differences were observed between ON and OFF medication in PD patients. Collectively, our data delineate the temporal dynamics of pre-saccadic attention in PD. Our findings suggest that aging reduces the build-up of pre-saccadic attention. This process is exacerbated by PD, but without a detectable influence of dopamine.


Talk 5, 6:30 pm

Fixational eye movements and visual acuity in patients with schizophrenia

Sanjana Kapisthalam^1,2 (), Howard Bi^1,2, Youjia Zhang¹, Ashley M. Clark^1,2, Judy L. Thompson¹, Martina Poletti^1,2, Brian P. Keane^1,2; ¹University of Rochester, ²Center for vision science

In non-clinical populations, visual acuity and oculomotor behavior at fixation are tightly linked. For the first time, we leveraged high-precision eye tracking to determine whether fixational eye movements are abnormal in schizophrenia and whether such abnormalities may contribute to reduced visual acuity. Visual acuity was first assessed with habitual correction in 14 patients and 13 age-matched healthy controls, none with ocular pathologies or worse than 20/20 Snellen acuity. To record eye movements with high precision, participants were asked to view stimuli without their habitual visual correction and were best corrected using a Badal Optometer. Subjects performed a 4AFC acuity task. Stimuli consisted of digits in Pelli’s font, whose size changed based on an adaptive staircase. Each trial began with a 400-ms blank screen followed by a 500-ms stimulus presentation. Although both groups averaged near 20/20, patients exhibited poorer Snellen acuity (20/20 vs. 20/16; p = 0.01, d = 1.06). Visual acuity thresholds measured with the adaptive procedure were slightly elevated but not significantly different in patients compared to controls (p = 0.3, Cohen’s d = 0.52). However, patients exhibited significantly greater between-subject variability in these thresholds (p = 0.03, Levene’s test). Our findings also revealed that fixation stability, quantified using the bivariate contour ellipse area, was reduced in patients (0.12± 0.03 deg^2 (patients) vs. 0.08±0.04 deg^2 (controls); p = 0.005, d = 0.98). Further, whereas healthy controls effectively suppressed microsaccades during stimulus presentation, patients did not and were characterized by a higher microsaccade rate (0.3±0.25 ms/s (controls) vs. 0.88±0.46 ms/s (patients), p = 0.003, d = 1.3). These findings show that patients with schizophrenia exhibit worse Snellen acuity, less stable fixation, and a higher rate of microsaccades, highlighting the possibility that these ocular abnormalities may contribute to an overall worse acuity.

NIH R21EY035001; Schmitt Program in Integrative Neuroscience (through Del Monte Institute at URMC)

Talk 6, 6:45 pm

Saccadic suppression in area MT/MTC is absent during simulated saccades in the visual input.

Amy Bucklaew¹, Shanna Coop², Jude Mitchell²; ¹University of Rochester, Neuroscience, ²University of Rochester, Brain and Cognitive Sciences

To maintain a stable percept of the world, visual information during saccades is suppressed to avoid retinal blur, a phenomenon termed saccadic suppression. Extra-retinal motor signals, potentially originating from efference copy feedback from frontal eye fields (FEF) or superior colliculus (SC), are thought to contribute to this suppression (Sommer & Wurtz, 2008). More recent studies show that saccadic suppression can start as early as the retina from wide-field rapid visual flow that simulates the visual input during saccades (Indrees et al., 2020; Baumann & Hafed 2024). In the present study we sought to clarify the involvement of suppression from the retina versus efference copy feedback at the level of visual cortex. Marmoset monkeys freely viewed either natural images, while in the dark, or while saccades were simulated by intermittently moving viewed natural images. We recorded from motion-selective areas MT and MTC. Previously, we reported that a subpopulation of neurons (18%), predominantly found in area MTC more so than MT, respond with early transients to saccades that are tuned for saccade direction and are faster than conventional visual response latencies (25-35ms vs 40-50ms). These short-latency responses could reflect an extra-retinal signal involved in driving saccadic suppression (Bucklaew et al., 2023). Here we tested if these short-latency responses could instead reflect a response to rapid full-field visual flow by simulating saccades with moving natural images and comparing it against saccades in free-viewing of natural images. We find that short-latency responses are absent with simulated saccades, as well as signs of the typical saccadic suppression in the rest of the population. Short-latency responses in the dark were highly attenuated, suggesting a modulatory rather than driving role for potential extra-retinal inputs. These results suggest extra-retinal feedback likely plays the key role in driving saccadic suppression in visual cortex.

Funding: AB, SC, and JFM from NIH EY030998, AB from NIH T32EY007125 and NIH F31EY035866

Talk 7, 7:00 pm

Dark contrasts are immune to saccadic suppression in the primary visual cortex

Wenbin Wu¹ (), Yue Yu¹, Tatiana Malevich¹, Matthias P. Baumann¹, Tong Zhang¹, Carlotta Trottenberg¹, Ziad M. Hafed¹; ¹University Tübingen

Saccade generation is accompanied by a dramatic reduction in perceptual sensitivity for perimovement stimulus onsets. Neuronal correlates of this phenomenon have been observed in multiple brain areas, including the retina, superior colliculus (SC), and primary visual cortex (V1). However, how each area specifically contributes to the perceptual effect itself remains unknown. Here, we were motivated by previous observations that perisaccadic perceptual detectability is similarly impaired for dark and bright stimuli, and that this is also true in the SC (Wu & Hafed, 2024). We asked whether these observations are universal (and thus observable in other brain areas), or whether they reveal a particular role for the SC in mediating them. We recorded from 248 SC neurons (three monkeys) and 325 V1 neurons (two monkeys). In each trial, a disc (0.51 deg radius; bright or dark) appeared within the recorded neurons’ response fields. We measured stimulus-evoked visual response strength as a function of stimulus onset time relative to microsaccades. In the SC, we replicated the earlier observations (Wu & Hafed, 2024) that suppression strength is similar for darks and brights. Surprisingly, this was not the case in V1: all of our dark contrasts were completely immune to saccadic suppression. Moreover, bright contrasts underwent weaker suppression than in the SC. These results suggest that perisaccadic perceptual detectability (similarly suppressed for darks and brights) is not mediated by V1. However, this does not mean a complete lack of V1 impact on perisaccadic vision. In five human subjects, we repeated the same experiments but now presented supra-threshold oriented bars (either dark or bright). The bars (at 20% contrast) were perisaccadically detectable in 100% of the trials. Nonetheless, orientation discrimination thresholds were still elevated, but only for the bright stimuli. Thus, there is a highly selective saccadic suppression of exclusively ON processing pathways in V1.