Searching for visuomotor matches vs mismatches biases confidence and visual sampling strategies, but not performance

Visuomotor self-other distinction relies on the comparison of forward predictions from one’s motor system with visual movement data. Previous work suggests that matching kinematics may be preferentially processed, but at the same time, visuomotor mismatches are known to capture attention. Here, participants were presented four virtual hands, each reflecting their actual hand movements, conveyed via a data glove, with a unique added time delay. Participants had to identify which of these hands moved most similarly (search for match) or most differently (search for mismatch) to their actual movements, under varying degrees of task difficulty. We found that the instruction to identify visuomotor matches vs mismatches significantly biased the participants’ confidence; i.e., participants exhibited overconfidence when searching for matches. Furthermore, eye tracking showed participants relied significantly more on serial sampling of the hands when searching for matches, but more on central fixation and peripheral vision when searching for mismatches. These biases were not universally reflected in detection performance. However, performance when searching for visuomotor matches was less strongly affected by task difficulty than when searching for mismatches.