Rapid eye movements do more than shift our gaze; experiments show that the distance the eyes travel can alter which features of a scene become most visible.
Study: What one sees depends on how far the eye has moved. Image Credit: antoniodiaz / Shutterstock
A recent study in the journal Nature Communications examined whether saccade amplitude, the distance the eye travels during a rapid gaze shift, acts as a visual filter that alters the relative visibility of coarse versus fine spatial detail, thereby actively shaping human perception beyond location selection.
Dynamic Encoding of Spatial Information in Vision
Sensory systems exhibit heightened sensitivity to input variability, leveraging these fluctuations to build sensory representations. Variations in sensory input provide critical cues about object motion and spatial structure. Notably, even stationary objects induce dynamic sensory signals during organismal movement, facilitating accurate localization and shape estimation.
Spatial information resulting from self-motion shapes perception across sensory modalities. In vision, although spatial coordinates are inherently mapped on the retina, temporal modulations from eye or head movements have primarily been examined in relation to depth perception. Yet, the robust responsiveness of the visual system to changes in input, coupled with incessant micro-movements of the eyes, even during fixation, indicates that dynamic input encoding is a core visual processing mechanism.
Ocular drift converts a static visual scene into a spatio-temporal luminance flow, supplying high-resolution spatial information within the temporal frequency band of peak neuronal sensitivity. Experimental findings demonstrate that the visual system uses these fluctuations to encode fine spatial details, as evidenced by impaired pattern discrimination when fixational motion is eliminated and by the observed coupling between fixational modulations and visual sensitivity.
Although research has predominantly focused on ocular drift, other eye movements may similarly engage dynamic space-time encoding. Lower spatial frequencies convey broad shapes and coarse structure, whereas higher spatial frequencies carry finer visual detail. Saccades, characterized by large, rapid retinal image displacements, generate substantial changes in input at low spatial frequencies, thereby enhancing sensitivity in this range compared with ocular drift.
Coarse-to-fine processing dynamics observed during the saccade-drift cycle indicate that the visual system integrates spatial information from both saccades and drifts. Nevertheless, a critical research gap persists: the precise effect of saccade amplitude on the spatial frequency profile of post-saccadic visual sensitivity remains unexamined.
Exploring the Role of Saccade Amplitude in Vision
The current study involved 21 participants (12 males, 9 females; ages 19–36), of whom 12 (7 males, 5 females) took part in the experiments reported in the study. Of these, 9 (5 males, 4 females) completed the instructed-saccade task (1° and 6° saccades), and 6 also participated in the 0.4° amplitude condition. Five (4 males, 1 female; ages 25–32) joined the hybrid-images experiment, including 2 who were also in the main group. Four among them further participated in both the simulated-saccade and broadband-stimuli experiments. Additional oculomotor data from 9 observers (ages 21–31), previously collected during free viewing of natural scenes, were used for spectral analyses and as a reference. All participants had 20/20 or better uncorrected visual acuity.
In the main experiments, participants viewed two superimposed, differently oriented gratings, with one serving as a fixed reference and the other as a probe at a lower or higher spatial frequency. In certain sessions, these gratings were replaced with oriented noise patterns or hybrid images that combined human and cartoon faces at different spatial frequencies. Right-eye movements were recorded at 1 kHz with a Dual Purkinje Image eye-tracker, and head stabilization ensured measurement accuracy.
Each session lasted about an hour and was organized into blocks of 100 trials. Calibration used a two-phase procedure for precise gaze localization. Trials began with fixation, followed by a contrast ramp and a saccade cue, with amplitude and direction varying by experiment. Both eye-movement and stimulus-visibility reports were collected. In simulated-saccade sessions, gratings moved across the screen based on pre-recorded saccade traces, isolating the contribution of retinal image motion from saccade-related extraretinal signals.
Saccade Amplitude Determines Spatial Frequency-Dependent Visual Sensitivity
Saccade amplitude strongly influenced visual sensitivity in a spatial-frequency-dependent manner by modulating saccade-induced luminance transients. Saccades, or rapid eye movements, generated luminance modulations determined by the spatial frequency of the stimulus and the amplitude of the saccade. Notably, saccade amplitude provided a separate and complementary form of selection, effectively filtering visual information by spatial frequency in addition to selecting it by location.
The results showed that larger saccades produced stronger luminance modulation within an amplitude-dependent band of low spatial frequencies, thereby enhancing the visibility of lower-spatial-frequency probes relative to smaller saccades. This enhancement remained robust across participants and experimental conditions, was reproduced with simulated saccade motion, and was unaffected by minor differences in saccade reaction time or direction.
Consistent with this retinal-signal analysis, perceptual measurements showed that a reduction in saccade amplitude attenuated post-saccadic visual sensitivity in an amplitude-dependent range of low spatial frequencies, with even small changes in saccade size considerably affecting visibility.
In forced-choice experiments, participants judged the visibility of two superimposed gratings after making saccades of different amplitudes. Results showed that larger saccades selectively improved the visibility of lower-spatial-frequency gratings, with contrast sensitivity increasing by 37% on average after 6° compared with 1° saccades. This outcome also generalized to more complex laboratory conditions, including broadband stimuli, unpredictably varying saccade amplitudes and directions, and naturalistic hybrid images, where larger saccades favored the low-frequency component, consistent with a coarser initial representation of the scene.
When the probe grating had a higher spatial frequency, experimental results indicated that 1° and 6° saccades produced virtually no difference in visibility. These findings support the view that the magnitude-dependent kinematic characteristics of saccades transform the visual world into a spatiotemporal flow, and that post-saccadic visibility closely follows theoretical predictions based on the spatial information conveyed by saccade transients within the temporal bandwidth of retinal sensitivity.
Experiments with very small saccades (0.4°) revealed a complementary pattern: the 0.4° saccade enhanced visibility of the higher-spatial-frequency probe compared with both the 1° and 6° saccades, while producing visibility of the lower-spatial-frequency probe similar to that of the 1° saccade. Together, these findings showed that the perceptual effects of saccades were governed by the interaction between saccade amplitude and stimulus spatial frequency.
Even small changes in eye-movement amplitude had significant consequences for visual perception, supporting the conclusion that saccades not only guided location-based selection but also filtered visual information by scale, actively shaping visual experience.
Conclusions
Saccades not only select visual information by location but also filter it by spatial frequency, actively shaping visual perception. Even small changes in saccade size can substantially alter visual sensitivity, underscoring the critical role of eye movement dynamics in visual processing.
The experiments involved relatively small groups of participants and were conducted largely under tightly controlled laboratory conditions, although the effects were consistent across observers and extended to more complex stimuli and viewing conditions.
Future research should explore whether amplitude-dependent filtering contributes to saccadic suppression or influences saccade planning under natural viewing conditions. The findings also raise the possibility that altered saccade dynamics could contribute to visual-processing differences in conditions associated with abnormal eye movements, although dedicated studies are needed to test these hypotheses.
Laboratory chemical exposure and poor diet increase odds of health symptoms in young scientists