Abstract
The amplitude of saccadic eye movements is adjusted by the oculomotor system. If a target is surreptitiously moved during saccades so that the eye lands beyond the target, a gradual decrease in the saccadic gain (saccade-size divided by target-displacement) ensues. What error signals guide this adaptation? We propose that visual attention helps keep track of the target across saccades, so that saccade accuracy can be determined by comparing the location of the fovea with the locus of attention. If attention plays this role, saccadic errors might be more apparent when the size of the attentional field is small than when it is large. We tested this by a conventional saccade adaptation paradigm in which the target stepped 8° and then stepped back by 30% during each saccade, while the subject did one of two psychophysical tasks demanding either a small or large attentional field. Specifically, the saccade target consisted of a large outer (8.5°) and a small inner (0.8°) ring, each with several breaks, rotating in opposite directions. After each saccade, the number of breaks transiently changed. Subjects reported the number of breaks present during this period in the attended ring. Saccadic gain was assessed both during adaptation and by comparing blocks of saccades before and after 250 adaptation trials using a spot target that disappeared upon saccade onset. After adaptation, saccadic gain had decreased four times as much if subjects attended to the small ring than to the large (small ring: −0.11, SD, 0.031; large ring: −0.027, SD, 0.038, paired t-test, p<0.01). Also, 4 of the 5 subjects had significantly negative gain slopes during the adaptation when attending to the small ring but none did when attending to the large ring. These findings argue that attention plays an important role in the adjustment of saccadic gain and suggest that the spatial scale of attention influences whether saccades are deemed to be accurate or in need of correction.