Prior expectations facilitate perceptual metacognition
In a new paper published in Consciousness and Cognition, Sackler Centre Ph.D. student Maxine Sherman shows how having prior expectations for whether a stimulus is likely to appear or not, can influence how good people's metacognitive judgements are - where metacognition here means how good people are at knowing whether their decisions are right or wrong. In a psychophysical experiment, she found that metacognitive decisions were more accurate when prior expectations of target presence/absence were congruent with participants responses. Importantly, the effect did not depend on attention. These findings shed more light on the influential 'Bayesian brain' or 'predictive processing' view of perception and cognition. Indeed, the paper includes a computational model which augments standard 'signal detection theory' models of perceptual decision with Bayesian components as a step towards integrating these distinct perspectives on the brain. The paper is co-authored by Maxine Sherman, Anil Seth, Adam Barrett, and Ryota Kanai.