Eye Movements: Neural dynamics, perceptual changes

Talk Session: Saturday, May 16, 2026, 8:15 – 9:45 am, Talk Room 1
Moderator: James Bisley, UCLA

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

Talk 1, 8:15 am, 21.11

The physiological correlates of conscious and unconscious visual and auditory processing: An MEG and eye-tracking study

Plyfaa Suwanamalik-Murphy¹ (), Sharif Kronemer¹, Victoria Gobo^1,2, Javier Gonzalez-Castillo¹, Amaia Benitez-Andonegui³, Anna Namyst³, Peter Bandettini^1,4; ¹Section on Functional Imaging Methods, Laboratory of Brain and Cognition, National Institute of Mental Health (NIMH), National Institutes of Health (NIH), Bethesda, Maryland (MD), United States of America (USA), ²Baylor College of Medicine, Houston, Texas, USA, ³MEG Core Facility, NIMH, NIH, Bethesda, MD, USA, ⁴Functional Magnetic Resonance Imaging Core Facility, NIMH, NIH, Bethesda, MD, USA

Distinguishing conscious from unconscious processing is a major challenge for uncovering the neural mechanisms of conscious perception and guiding the assessment of conscious level in patient groups with disordered consciousness. Here, we explore neural and non-neural physiological metrics and their signal properties that may differentiate conscious and unconscious sensory processing in vision and audition. We explored eye metrics (pupil size, blinking, microsaccades) and cortical dynamics measured with magnetoencephalography (MEG). Two major signal characteristics were assessed: (1) magnitude and (2) timing (e.g., the latency of change from stimulus onset and duration of response). Healthy adult participants performed a visual (N = 33) and auditory (N = 37) stimulus detection task with simultaneous eye-tracking, pupillometry, and MEG. Sub-, near-, and supra-threshold stimuli were presented on the left or right side of a screen (face images) or in the left or right ear (tones). Participants indicated when they perceived a stimulus and its location (on-screen or in which ear). Our results indicate strong correspondence across behavioral, eye, and MEG responses and between visual and auditory modalities. As stimulus intensity increased, perception rate increased, and reaction time decreased. Perceived stimuli elicited increased pupil size and blink rate and decreased microsaccade rate in auditory and visual modalities, relative to unperceived stimuli. Eye metric responses changed systematically with stimulus intensity regardless of conscious perception: as intensity increased, response amplitudes increased, and latencies decreased. Similarly, for visual and auditory stimuli, MEG activity showed larger magnitude and more prolonged responses for perceived versus unperceived stimuli. However, regardless of perception, MEG signal magnitude and latency were similarly modulated by stimulus intensity in occipital (visual) and temporal (auditory) scalp regions: larger and more rapid MEG responses as stimulus intensity increased. These findings help clarify perception-dependent and independent components of eye and cortical signals in conscious and unconscious vision and audition.

NIH Intramural Research Program ZIAMH002783

Talk 2, 8:30 am, 21.12

Frequency selective post-microsaccadic perceptual enhancements in the fovea

Zoe Stearns^1,2, Krishnamachari Prahalad^1,2, Martina Poletti^1,2; ¹Center for Visual Science, ²University of Rochester

Microsaccades compensate for non-uniform sensitivity within the foveola by precisely redirecting the preferred locus of fixation (PLF) and by enhancing fine spatial vision at their target location before their onset. Yet, whether and how visual perception is also modulated upon microsaccade landing remains unclear. To address this, we tested post-microsaccadic visual discrimination of different spatial frequencies at the landing site and further away. Eye movements were recorded with a high-resolution Dual-Purkinje Image eye tracker as observers freely viewed an array of five gabors spanning 2.5° (0.5° center-to-center) at the center of the display. At variable times all the gabors changed phase and one also changed orientation (±45°). Subjects (N=10) reported the direction of the brief orientation change (2AFC). Stimulus contrast was adjusted so that performance was at threshold when the orientation-change occurred during fixation at the PLF. Subjects spontaneously performed frequent microsaccades during free viewing. Visual discrimination performance was then evaluated at baseline when subjects were maintaining fixation and 0-130ms after microsaccade landing. Performance was assessed both for stimuli near (within 15 arcmin) and far (20–40 arcmin) from the PLF and for low (4 cpd) and high (16 cpd) spatial frequencies. Visual discrimination was enhanced relative to baseline immediately after microsaccade landing (rm-ANOVA: F(1,9)=8.19, p=0.019). This enhancement was spatially localized around the PLF (F(1,9)=11.83, p=0.007) and selective for higher spatial frequencies (F(1,9)=12.69, p=0.006). For high spatial frequencies, the post-microsaccadic gain increased over time, reaching significance ~40 ms upon landing, whereas the performance for low spatial frequencies remained constant in the period following microsaccade landing. These results show that microsaccades are associated with post-microsaccadic perceptual enhancements, these modulations are spatial-frequency selective and are localized at the microsaccade landing site, indicating that they are likely driven by extraretinal signals rather than temporal modulations of the visual input brought by the saccadic transient.

funded by Meta and NIH grant EY001319 to the Center for Visual Science.

Talk 3, 8:45 am, 21.13

Motor priority accelerates recovery from global saccadic inhibition

Paul Schmitthäuser^1,2 (), Martin Rolfs¹, Nina M. Hanning¹; ¹Humboldt-Universität zu Berlin, ²Freie Universität Berlin

We generate saccadic eye movements at a baseline rate of approximately 2 to 4 per second. Sudden visual transients reliably interrupt this stream of eye movements, producing a sharp suppression of saccade initiation—a phenomenon known as “saccadic inhibition”—that is followed by a rebound that briefly exceeds baseline. Although saccadic inhibition has been demonstrated in many visual tasks, it remains elusive how this inhibition interacts with the unfolding dynamics of saccade planning in naturalistic free viewing, and whether it depends on the spatial congruency between a transient and the current motor plan. Participants performed an unrestricted search task in which they freely explored an array (r=~14°) of 19 differently colored tiles until they found a hidden target. Tile colors signaled target probabilities and encouraged systematic, predictable gaze patterns. Brief luminance probes (8.3 ms) were flashed either at the current fixation location, the predicted saccade target, or a non-target location. The natural variability in probe timing relative to saccade onset allowed us to probe different phases of saccade preparation. Saccadic inhibition followed probe onsets at any location, regardless of congruency with the saccade target. Critically, however, recovery was faster when probes coincided with the upcoming saccade target than when they appeared at other locations. These results suggest that visual transients interact with naturalistic visual exploration in two temporally distinct ways: an initial, global pause that interrupts saccade initiation, followed by a spatially selective recovery that is facilitated when the transient is congruent with the current saccade plan. This pattern aligns with an architecture in which global inhibition temporarily halts saccade execution, while spatially selective motor priority is preserved and shapes the speed of recovery.

This research was supported by the European Research Council (ERC) under the European Union’s Horizon Europe research and innovation programme (project 101165343 – FREE-VIEW to NMH).

Talk 4, 9:00 am, 21.14

Top-Down Motivation Both Increases and Decreases Feature Interference Following a Saccade

Tzu-Yao Chiu¹ (), Julie D. Golomb¹; ¹The Ohio State University

Spatial attention is rarely static, but instead highly dynamic in real life. Recent studies have revealed that shifts of attention and eye movements incur consequences for ongoing feature perception, such as systematic feature-binding (swapping and mixing) errors. Here we investigated if these consequences are sensitive to top-down control, such that they could be flexibly mitigated when participants are motivated to improve their performance, or if feature interference reflects a more automatic byproduct of attentional remapping mechanisms following a saccade. In a pre-registered experiment, 40 healthy adults reported the color of an item appearing at a pre-cued spatiotopic (world-centered) location. Following the cue, participants made a saccade, after which an array of four colored items briefly appeared. One of the colored items appeared at the cued spatiotopic target location, one at the cue’s retinotopic (eye-centered) location, and two at control non-target locations. Crucially, across blocks we provided two different levels of performance-contingent reward (monetary incentives for precise spatiotopic target reports) to induce differences in top-down motivation. Compared to baseline lower-reward blocks, during higher performance-contingent reward (“bonus”) blocks, participants showed improved general performance and reduced subtle retinotopic mixing errors immediately after the saccade. However, we found the opposite effect on retinotopic swapping errors, with higher reward/motivation associated with increased retinotopic swapping errors early after the saccade, indicating that retinotopic swapping errors may reflect a structural limitation of the attentional remapping process. Overall, these results uncover the role of motivation in distinct attentional mechanisms underlying the remapping of spatial attention across saccades, and provide novel insights into the role of top-down motivation in dynamic spatial attention and visual perception more generally.

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

Talk 5, 9:15 am, 21.15

Questioning the functional validity of pre-saccadic remapping in LIP

James Bisley¹, Yelda Alkan¹, Alanna Morris¹; ¹UCLA

Despite constantly shifting our gaze, our perception remains stable. It has been proposed that predictive remapping of receptive fields (RFs) in the lateral intraparietal area (LIP) may play a role in maintaining this stability. A number of studies have shown that remapped responses can begin before the saccade: pre-saccadic remapping. To improve our understanding of remapping, we investigated how the remapped response relates to saccade length. We recorded LIP activity in 2 animals performing a saccade task in which a task irrelevant probe appeared in the post-saccadic RF before the animals made a 7, 14 or 21 degree saccade. RFs were mapped using flashed stimuli and a memory guided saccade task. Despite the care taken to map the RFs, we found that some units responded significantly to the probe appearing in the post-saccadic RF prior to the 7 deg saccade. For the main experiment, these conditions were excluded because the probe was not presented outside the RF. Surprisingly, we found that pre-saccadic remapping only occurred in the excluded condition. No units for which the probe was presented clearly outside of the RF showed pre-saccadic remapping, and units that had pre-saccadic responses only had so for the excluded condition – none showed pre-saccadic remapping in the 14 and 21 deg saccade conditions. The few prior studies that have shown pre-saccadic remapping in LIP often do not describe the lengths of saccades nor provide detailed receptive field mapping methods, or only use RF “hot spots”. And remapping studies looking at other cortical areas either have not aligned data by saccade onset or have used slow infrared eye tracking systems and large analysis windows that do not allow for accurate timing estimates. We propose that pre-saccadic responses may be a result of within neuron processing rather than a fundamental feature of LIP remapping.

This work was supported by the National Eye Institute (R01 EY032863).

Talk 6, 9:30 am, 21.161

Object files are lost at saccadic speeds

Melis Ince^1,2 (), Carolin Hübner³, Martin Rolfs^1,2; ¹Humboldt-Universität zu Berlin, ²Berlin School of Mind and Brain, ³Technische Universität Chemnitz

Object motion that follows the velocity-duration-amplitude relationship of saccades can support perceived object continuity even when the motion itself is not consciously perceived (Ince et al., 2025). Here we tested whether object files—temporary representations of objects and their features (Kahneman et al., 1992)—survive motion at saccadic speed. In an adapted object-reviewing paradigm (Sasi et al., 2023), trials started with a brief preview of two digits displayed inside two Gabor patches. The patches then blanked and reappeared at the same (stationary) or new locations (apparent motion), or moved smoothly to their new locations (continuous motion). The duration of the blank/motion phase was set to match saccade kinematics for the movement distance (6 dva) used in the motion conditions (mean velocity: 39 ms, 154.3 dva/s; peak velocity: 25 ms, 239.4 dva/s), or a previously used lower velocity (300 ms, 20 dva/s; Sasi et al., 2023). After the blanking/motion, a target digit appeared in one of the two Gabors, which participants (n = 19) had to rapidly identify. This target digit was either one of the previewed digits (from the same or different object) or a new digit. At long durations (slow velocity), we replicated Object-Specific Preview Benefits (OSPBs)—shorter response times for digits previously shown in the same compared to a different object (Sasi et al., 2023). Consistent with object-file theory, these OSPBs were evident in the static and continuous conditions, but not in apparent motion, where the same and different objects could not be distinguished. OSPBs were absent for motion at saccadic speeds (mean and peak velocities). These results suggest that—although motion at saccadic speed can disambiguate object correspondence through spatiotemporal continuity (Ince et al., 2025)—object-file representations (that include object features) do not survive these rapid retinal position changes.

This research was supported by the European Research Council under the European Union’s Horizon 2020 research and innovation program (Grant 865715), and the Deutsche Forschungsgemeinschaft (grants RO3579/8-1, RO3579/10-1).