Owen Vegas¹, Sydney Trowbridge², Silvia Lopez-Guzman², Michael Grubb¹; ¹Trinity College, Hartford CT, ²Unit on Computational Decision Neuroscience, National Institute of Mental Health, National Institutes of Health

Cognitive training interventions like meditation practice and video game playing have been shown to increase attentional control, but the impact of metacognitive training has yet to be fully examined. We designed an experiment to test the hypothesis that making metacognitive judgements about perceptual decisions induces state changes in attentional control. 186 participants completed an attentional blink task, which required reporting the identity of target 1 and the presence/absence of target 2 at varying temporal lags. Next, participants completed a simple orientation discrimination task: one of four Gabors was rotated clockwise or counterclockwise from vertical. Depending on group assignment, participants then reported their confidence in their orientation decision (1-4 scale), reported the location of the non-vertical stimulus (1 of 4 locations), or made no secondary report. All participants then redid the attentional blink task. We analyzed the data with a signal detection theoretic, linear mixed effects model fit with Hierarchical Bayesian estimation, designed to quantify the change in sensitivity and response criterion from pretest to posttest in each group. Sensitivity quantifies participants’ ability to correctly identify the presence of target 2, reflecting attentional control. Criterion quantifies participants’ biases to report present or absent. We found that sensitivity significantly increased in the confidence group from pretest to posttest, as well as in the single judgment group. We conclude that these improvements were likely due to practice effects. Notably, the localization group did not experience these improvements. Interestingly, metacognitive practice impacted response criterion: the confidence group became more conservative in the posttest, while the single judgment and localization groups became more liberal. In short, repeated self-assessment of one’s confidence in an orientation discrimination task induced conservative response biases in a subsequent detection task, highlighting the impact of metacognitive practice on visual decision making.

Acknowledgements: This research was supported (in part) by the Intramural Research Program of the NIMH - ZIAMH002988, and by NSF-2141860 CAREER Award to MAG.