A neural basis of probabilistic computation in visual cortex

Edgar Y. Walker, R. James Cotton, Wei Ji Ma, Andreas S. Tolias

Research output: Contribution to journalArticle

Abstract

Bayesian models of behavior suggest that organisms represent uncertainty associated with sensory variables. However, the neural code of uncertainty remains elusive. A central hypothesis is that uncertainty is encoded in the population activity of cortical neurons in the form of likelihood functions. We tested this hypothesis by simultaneously recording population activity from primate visual cortex during a visual categorization task in which trial-to-trial uncertainty about stimulus orientation was relevant for the decision. We decoded the likelihood function from the trial-to-trial population activity and found that it predicted decisions better than a point estimate of orientation. This remained true when we conditioned on the true orientation, suggesting that internal fluctuations in neural activity drive behaviorally meaningful variations in the likelihood function. Our results establish the role of population-encoded likelihood functions in mediating behavior and provide a neural underpinning for Bayesian models of perception.

Original languageEnglish (US)
Pages (from-to)122-129
Number of pages8
JournalNature Neuroscience
Volume23
Issue number1
DOIs
StatePublished - Jan 1 2020

Fingerprint

Likelihood Functions
Visual Cortex
Uncertainty
Population
Primates
Neurons

ASJC Scopus subject areas

  • Neuroscience(all)

Cite this

A neural basis of probabilistic computation in visual cortex. / Walker, Edgar Y.; Cotton, R. James; Ma, Wei Ji; Tolias, Andreas S.

In: Nature Neuroscience, Vol. 23, No. 1, 01.01.2020, p. 122-129.

Research output: Contribution to journalArticle

Walker, Edgar Y. ; Cotton, R. James ; Ma, Wei Ji ; Tolias, Andreas S. / A neural basis of probabilistic computation in visual cortex. In: Nature Neuroscience. 2020 ; Vol. 23, No. 1. pp. 122-129.
@article{8126561cab82478fa9d164c463839e96,
title = "A neural basis of probabilistic computation in visual cortex",
abstract = "Bayesian models of behavior suggest that organisms represent uncertainty associated with sensory variables. However, the neural code of uncertainty remains elusive. A central hypothesis is that uncertainty is encoded in the population activity of cortical neurons in the form of likelihood functions. We tested this hypothesis by simultaneously recording population activity from primate visual cortex during a visual categorization task in which trial-to-trial uncertainty about stimulus orientation was relevant for the decision. We decoded the likelihood function from the trial-to-trial population activity and found that it predicted decisions better than a point estimate of orientation. This remained true when we conditioned on the true orientation, suggesting that internal fluctuations in neural activity drive behaviorally meaningful variations in the likelihood function. Our results establish the role of population-encoded likelihood functions in mediating behavior and provide a neural underpinning for Bayesian models of perception.",
author = "Walker, {Edgar Y.} and Cotton, {R. James} and Ma, {Wei Ji} and Tolias, {Andreas S.}",
year = "2020",
month = "1",
day = "1",
doi = "10.1038/s41593-019-0554-5",
language = "English (US)",
volume = "23",
pages = "122--129",
journal = "Nature Neuroscience",
issn = "1097-6256",
publisher = "Nature Publishing Group",
number = "1",

}

TY - JOUR

T1 - A neural basis of probabilistic computation in visual cortex

AU - Walker, Edgar Y.

AU - Cotton, R. James

AU - Ma, Wei Ji

AU - Tolias, Andreas S.

PY - 2020/1/1

Y1 - 2020/1/1

N2 - Bayesian models of behavior suggest that organisms represent uncertainty associated with sensory variables. However, the neural code of uncertainty remains elusive. A central hypothesis is that uncertainty is encoded in the population activity of cortical neurons in the form of likelihood functions. We tested this hypothesis by simultaneously recording population activity from primate visual cortex during a visual categorization task in which trial-to-trial uncertainty about stimulus orientation was relevant for the decision. We decoded the likelihood function from the trial-to-trial population activity and found that it predicted decisions better than a point estimate of orientation. This remained true when we conditioned on the true orientation, suggesting that internal fluctuations in neural activity drive behaviorally meaningful variations in the likelihood function. Our results establish the role of population-encoded likelihood functions in mediating behavior and provide a neural underpinning for Bayesian models of perception.

AB - Bayesian models of behavior suggest that organisms represent uncertainty associated with sensory variables. However, the neural code of uncertainty remains elusive. A central hypothesis is that uncertainty is encoded in the population activity of cortical neurons in the form of likelihood functions. We tested this hypothesis by simultaneously recording population activity from primate visual cortex during a visual categorization task in which trial-to-trial uncertainty about stimulus orientation was relevant for the decision. We decoded the likelihood function from the trial-to-trial population activity and found that it predicted decisions better than a point estimate of orientation. This remained true when we conditioned on the true orientation, suggesting that internal fluctuations in neural activity drive behaviorally meaningful variations in the likelihood function. Our results establish the role of population-encoded likelihood functions in mediating behavior and provide a neural underpinning for Bayesian models of perception.

UR - http://www.scopus.com/inward/record.url?scp=85077062996&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85077062996&partnerID=8YFLogxK

U2 - 10.1038/s41593-019-0554-5

DO - 10.1038/s41593-019-0554-5

M3 - Article

C2 - 31873286

AN - SCOPUS:85077062996

VL - 23

SP - 122

EP - 129

JO - Nature Neuroscience

JF - Nature Neuroscience

SN - 1097-6256

IS - 1

ER -