Limits...
Suppression and reversal of motion perception around the time of the saccade

View Article: PubMed Central

ABSTRACT

We make fast, “saccadic” eye movements to capture finely resolved foveal snapshots of the world but these saccades cause motion artefacts. The artefacts go unnoticed, perhaps because the brain suppresses them through subcortical oculomotor signals feeding back into visual cortex. Opposing views, however, claim that passive mechanisms suffice: saccadic shearing forces might render the retina insensitive to the artefacts or post-saccadic snapshots might mask them before they enter consciousness. Crucially, only active suppression could explain perceptual changes that precede saccades but existing evidence for presaccadic misperception are ill-suited for addressing this issue: Previous studies have found misperceptions of space for objects briefly flashed before saccades, but perhaps only because observers confused the timing of flashes and saccades before they could be tested (“postdiction”), and presaccadic motion perception might have appeared to decline because motion stimuli persisted past eye movement onset. Here we addressed these concerns using briefly flashed two-frame animations (50 ms) to probe people’s motion sensitivity during and around saccades. We found that sensitivity declined before saccade onset, even when the probe appeared entirely outside the saccade, and this sensitivity decline was present for motion in every direction relative to saccade, ruling out problems with postdiction. Intriguingly, brief periods during the saccade produced negative sensitivity as if motion was reversed, arguably due to postsaccadic enhancement. These data suggest that motion perception is minimized during saccades through active suppression, complementing neurophysiological findings of colliculo-pulvinar projections that suppress the cortical middle temporal area around the time of the saccade.

No MeSH data available.


Motion sensitivity when a saccade is imminent. (A) Schematic illustrating the epochs for motion onset that fall outside of the perisaccadic interval (solid orange region) and the epoch where saccades are imminent (orange striped region). (B) Sensitivity to parallel and orthogonal motion probes outside of the perisaccadic interval and when saccades are imminent. Error bars represent the standard error.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4628122&req=5

Figure 4: Motion sensitivity when a saccade is imminent. (A) Schematic illustrating the epochs for motion onset that fall outside of the perisaccadic interval (solid orange region) and the epoch where saccades are imminent (orange striped region). (B) Sensitivity to parallel and orthogonal motion probes outside of the perisaccadic interval and when saccades are imminent. Error bars represent the standard error.

Mentions: To avoid these problems, we sidestepped the fitted inverted Gaussians and reanalyzed the data focusing on two particular time windows. We defined an “imminent” time window or epoch that included motion probes presented between 65 ms and 25 ms before saccade onset (Figure 4A, orange-stripe region), that is, the window included motion probes that were presented just before saccade onset but excluded probes that reached into the saccade. An “outside” window collected all trials with motion probes presented well before (300 ms to 100 ms) and well after (135 ms to 300 ms) saccade onset (Figure 4A, orange-solid regions). Sensitivity values for these times submitted to a 2-way repeated-measures ANOVA produced a main effect of “Epoch” (imminent/outside; F(1, 7) = 16.67, p = 0.005; Figure 4B). This shows that motion sensitivity started to decline before saccade onset. Furthermore, the ANOVA yielded a main effect of “Motion plane” (F(1, 7) = 9.08, p = 0.020), consistent with the extrasaccadic differences between parallel vs. orthogonal sensitivity mentioned earlier (Figure 3B, left). Interestingly, there was no interaction between “Epoch” and “Motion plane” (F(1, 7) = 0.58, p = 0.471), indicative of two independent mechanisms with additive effects on sensitivity.


Suppression and reversal of motion perception around the time of the saccade
Motion sensitivity when a saccade is imminent. (A) Schematic illustrating the epochs for motion onset that fall outside of the perisaccadic interval (solid orange region) and the epoch where saccades are imminent (orange striped region). (B) Sensitivity to parallel and orthogonal motion probes outside of the perisaccadic interval and when saccades are imminent. Error bars represent the standard error.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4628122&req=5

Figure 4: Motion sensitivity when a saccade is imminent. (A) Schematic illustrating the epochs for motion onset that fall outside of the perisaccadic interval (solid orange region) and the epoch where saccades are imminent (orange striped region). (B) Sensitivity to parallel and orthogonal motion probes outside of the perisaccadic interval and when saccades are imminent. Error bars represent the standard error.
Mentions: To avoid these problems, we sidestepped the fitted inverted Gaussians and reanalyzed the data focusing on two particular time windows. We defined an “imminent” time window or epoch that included motion probes presented between 65 ms and 25 ms before saccade onset (Figure 4A, orange-stripe region), that is, the window included motion probes that were presented just before saccade onset but excluded probes that reached into the saccade. An “outside” window collected all trials with motion probes presented well before (300 ms to 100 ms) and well after (135 ms to 300 ms) saccade onset (Figure 4A, orange-solid regions). Sensitivity values for these times submitted to a 2-way repeated-measures ANOVA produced a main effect of “Epoch” (imminent/outside; F(1, 7) = 16.67, p = 0.005; Figure 4B). This shows that motion sensitivity started to decline before saccade onset. Furthermore, the ANOVA yielded a main effect of “Motion plane” (F(1, 7) = 9.08, p = 0.020), consistent with the extrasaccadic differences between parallel vs. orthogonal sensitivity mentioned earlier (Figure 3B, left). Interestingly, there was no interaction between “Epoch” and “Motion plane” (F(1, 7) = 0.58, p = 0.471), indicative of two independent mechanisms with additive effects on sensitivity.

View Article: PubMed Central

ABSTRACT

We make fast, “saccadic” eye movements to capture finely resolved foveal snapshots of the world but these saccades cause motion artefacts. The artefacts go unnoticed, perhaps because the brain suppresses them through subcortical oculomotor signals feeding back into visual cortex. Opposing views, however, claim that passive mechanisms suffice: saccadic shearing forces might render the retina insensitive to the artefacts or post-saccadic snapshots might mask them before they enter consciousness. Crucially, only active suppression could explain perceptual changes that precede saccades but existing evidence for presaccadic misperception are ill-suited for addressing this issue: Previous studies have found misperceptions of space for objects briefly flashed before saccades, but perhaps only because observers confused the timing of flashes and saccades before they could be tested (“postdiction”), and presaccadic motion perception might have appeared to decline because motion stimuli persisted past eye movement onset. Here we addressed these concerns using briefly flashed two-frame animations (50 ms) to probe people’s motion sensitivity during and around saccades. We found that sensitivity declined before saccade onset, even when the probe appeared entirely outside the saccade, and this sensitivity decline was present for motion in every direction relative to saccade, ruling out problems with postdiction. Intriguingly, brief periods during the saccade produced negative sensitivity as if motion was reversed, arguably due to postsaccadic enhancement. These data suggest that motion perception is minimized during saccades through active suppression, complementing neurophysiological findings of colliculo-pulvinar projections that suppress the cortical middle temporal area around the time of the saccade.

No MeSH data available.