Abstract
During saccadic eye movements visual sensitivity decreases due to saccadic suppression1, which degrades spatial information of the target location. This implies that the visual system does not have precise information of target location during saccades. However, previous work showed that sensitivity to displacement during saccades is improved by blanking the target prior to displacement2. These results are consistent with the idea that the visual system assumes that the environment is stable unless there is "sufficient" evidence to the contrary. Alternatively, the visual system could be using both retinal and extra-retinal information in an optimal way3. If this is true, perceptual stability, measured as sensitivity to displacement, should correlate with oculomotor stability such that reduced saccadic accuracy and precision should lead to poor perceptual performance. In Experiment 1, each trial began with a fixation target. After 1-1.5s, the target jumped either to the right or left by 8 degrees. This jump served as a "go" signal for the observer to make an eye movement to this peripheral target. Contingent upon saccade onset, the target was then displaced with various amounts right or left. On 'blank' trials, the target disappeared after saccade initiation for 200msec. Observers reported the direction of target displacement. In Experiment 2, we measured fixation stability while observers viewed an empty screen or a fixation target on an otherwise blank screen. Psychometric functions were fit to observers' displacement direction responses. We computed correlations between the bias and slope parameters of the psychometric functions and oculomotor metrics such as fixation stability, saccade accuracy, and saccade precision. We found a significant correlation between the slope of psychometric functions and both signed saccade accuracy and precision, only in the 'blank' condition. This suggests that retinal and extra-retinal information can be used optimally if there is sufficient evidence to overcome the stability assumption.
Meeting abstract presented at VSS 2018