Abstract
How does orienting of attention in space affect one’s ability to evaluate one’s performance (i.e., metacognition)? Previous work has considered cases in which spatial attention is already fully deployed, but less is known about metacognition during attentional deployment. Here, we investigated how the timing of attention affected metacognitive ability. To probe exogenous and endogenous visuo-spatial attention, we adapted a “Wundt clocks” paradigm. This design builds on the fact that attention has been shown to alter the latency between objective and perceived events (i.e., “flash-lag” effect). Participants looked at 6 clocks at a fixed eccentricity that rotated at a fixed speed but different phases. At a random time, one of the clocks was either cued peripherally (exogenous) or centrally (endogenous), and when the clocks stopped, participants were requested to report the hand position at cue onset. The moment of attentional orienting was manipulated using a “pre-cue” condition, such that attention could have already been deployed at cued location or has yet to be deployed. After two trials, participants chose the one they felt more confident to be correct. The average reported times were delayed in accordance with exogenous/endogenous attention. Surprisingly, confidence was not correlated to these attention-induced delays. However, confidence judgments correlated with the relative error between each trial in the pair, suggesting that participants were able to estimate their internal deviation at the trial level. Importantly, exogenous/endogenous orienting of attention reduced this confidence-error relation compared to the pre-cue condition. To investigate this behavior, a second task required participants to make a speeded response followed by a confidence evaluation of their own reaction times. Metacognitive abilities in both tasks were correlated. These results suggest that orienting exogenous/endogenous attention in space alters metacognitive ability, possibly through domain-general impairment of internal time monitoring.