Abstract
The accuracy of saccadic eye movements is preserved by saccadic adaptation. It is commonly believed that saccadic adaptation relies on visual prediction error, i.e. a mismatch between the expected and actual sensory consequences of the saccade. However, recent research indicates that saccadic adaptation is driven by postdicted motor error, i.e. the post-saccadic error is used to update the representation of visual space and the estimated pre-saccadic target position. In our current study, we investigated whether oculomotor behavior can be adapted based on post-saccadic information only. We measured eye movements and localization judgements of human subjects in a novel visuomotor learning task during which only post-saccadic information about the target was available. During feedback trials, participants performed a saccade to the position where they expected the target to appear. Once saccade onset was detected, the target stimulus appeared. Each feedback trial was followed alternatively by a pre- or post-saccadic localization trial. During post-saccadic localization trials, participants performed a saccade to the position where they expected the target to appear and then moved a cursor to the position where they expected the target. During pre-saccadic localization trials, subjects maintained fixation and simultaneously moved the cursor to the position where they expected the target. Following a baseline period with stable target position, the target was stepped 0.03 deg each learning trial. Inward and outward target displacement was measured in separate conditions (n = 16). We found that participants adjusted their saccade amplitude to the changing target position. Further, both pre- and post-saccadic localization judgements shifted in direction of the target displacement. Our results indicate that post-saccadic information is sufficient to induce changes to saccade amplitude and target localization, possibly reflecting changes to the target representation driven by postdicted motor error.