Abstract
Saccadic adaptation is an adaptive learning mechanism used to ensure saccade accuracy. Previous research demonstrated that saccadic adaptation is driven by post-saccadic errors, which are most effective directly after saccade offset (Shafer, Noto, & Fuchs, 2000). Here, we test the hypothesis that error signals driving saccadic adaptation can also be evaluated in visual working memory – long after saccade offset. Participants were instructed to execute a saccade and judge the orientation of one out of two appearing Gabor patches. Before the saccade, only a perceptually-irrelevant saccade target was displayed at a horizontal eccentricity of 10°. At saccade onset, two Gabor patches appeared for 150 ms near the saccade target, vertically displaced from each other by 4°. Two seconds after the Gabors’ offset, an auditory cue signaled which Gabor had to be judged subsequently. Which Gabor was cued varied randomly across trials. One and a half seconds after the auditory cue, a reference Gabor appeared in-between the locations of the previously presented Gabors and participants had to judge if the orientation of the reference and the cued Gabor matched. Vertical saccade amplitudes showed robust trial-to-trial changes in the direction of the previously cued Gabor. The magnitude of adaptation was comparable to a control experiment, in which only a single, task-relevant Gabor was shown at random locations directly after saccade onset. Our findings demonstrate that error signals can be evaluated in visual working memory, long after the offset of visual targets and the saccade. Either errors are calculated directly after the saccade, and selected subsequently in memory, or competing target locations are stored in memory and errors are computed on the memorized locations. Both cases might be accomplished by orienting attention in visual working memory (Griffin & Nobre, 2003).
Acknowledgement: This project was funded by the SFB/TRR 135 and the International Research Training Group, IRTG 1901,