Skip to content

A leaky evidence accumulation process for perceptual experience.

Michael Pereira, Denis Perrin, Nathan Faivre

Trends in cognitive sciences June 1, 2022 DOI: 10.1016/j.tics.2022.03.003 via PubMed

Summary

AI-generated from the abstract

Perceptual experiences are not static but fluctuate as a leaky evidence accumulation process (LEAP), where confidence corresponds to the peak accumulated evidence. This proposal extends evidence accumulation models from decision-making to explain the dynamics of conscious perception itself, suggesting that percepts fade in and out of consciousness independent of tasks. The account offers a mechanistic understanding of phenomenal qualities like intensity and duration, moving beyond simple detection. Limitations and implications for neural correlates of consciousness are discussed.

Study at a glance

Characteristics Theoretical or philosophical paper Peer reviewed
Keywords Decision-making Electrophysiology Evidence accumulation Perceptual consciousness Perceptual monitoring
Key finding Percepts fade in and out of consciousness according to the dynamics of a leaky evidence accumulation process, with confidence reflecting maximal accumulated evidence.

Abstract

The neural correlates supporting our perceptual experience of the world remain largely unknown. Recent studies have shown how stimulus detection and related confidence involve evidence accumulation (EA) processes similar to those involved in perceptual decision-making. Here, we propose that independently from any tasks, percepts are not static but fade in and out of consciousness according to the dynamics of a leaky evidence accumulation process (LEAP), and that confidence corresponds to the maximal evidence accumulated by this process. We discuss the implications and limitations of our proposal, assess how it may qualify as a neural correlate of consciousness, and illustrate how it brings us closer to a mechanistic understanding of phenomenal aspects of perceptual experience like intensity and duration, beyond mere detection.

Comments

No comments yet.

Log in to comment