The Perceptual Basis of Confidence

Project Awarded: $20,685

A science of consciousness hinges on our ability to measure subjective perceptual experience. One candidate measure of subjective perception is confidence. When making a perceptual decision—for example, about whether a person in the distance is your friend—observers might be very confident when the decision is easy (e.g. the person is fairly close), but much less confident when the decision is difficult (e.g. the person is quite far). Our goal here is to determine how confidence about the orientation of a visual stimulus is computed in the human brain. Orientation tasks provide a tractable way to study perceptual decision making, as much is already known about how these tasks are represented in the brain. To investigate confidence, we will leverage a recently developed approach to dissociate confidence from accuracy known as “positive evidence bias”. Subjective confidence typically correlates with objective accuracy—an observer will usually identify a nearby person more accurately as well as be more confident about the decision—so dissociating these two types of measures is critical for determining the neural mechanisms that underlie confidence, specifically. In positive evidence bias, confidence differs in two experimental conditions even though accuracy is the same, creating in normal observers a situation mirroring the rare neurological condition known as “blindsight”. In this project, we will combine computational modeling, magnetoencephelography measurements, and behavioral experiments in humans to test mechanisms underlying the computation of confidence. This integrated approach is designed to yield a theoretical account of objective decision and subjective confidence reports across a wide range of orientation judgment tasks, with applicability to perceptual decision-making more broadly. It will, moreover, provide a body of evidence to guide our interpretation of confidence measures, favoring either a perceptual or post-perceptual (i.e. decision-based or metacognitive) basis for confidence. These findings will thus inform our understanding of subjective perception and how best to measure it as well as philosophical debates about the nature of subjective perception, the perception/cognition divide, and the epistemological status of perceptual states and states of confidence.

Team members:

 Rachel Denison, PhD. Post-doctoral fellow, Department of Psychology and Neural Science, New York University

Rachel Denison, PhD. Post-doctoral fellow, Department of Psychology and Neural Science, New York University

 Jason Samaha, PhD student, Department of Psychology, University of Wisconsin- Madison

Jason Samaha, PhD student, Department of Psychology, University of Wisconsin- Madison

 Tony Cheng, PhD student, Department of Philosophy, University College London

Tony Cheng, PhD student, Department of Philosophy, University College London