Trial-to-trial, uncertainty-based adjustment of decision boundaries in visual categorization

Ahmad T. Qamar, R. James Cotton, Ryan G. George, Jeffrey M. Beck, Eugenia Prezhdo, Allison Laudano, Andreas S. Tolias, Wei Ji Ma

Research output: Contribution to journalArticle

Abstract

Categorization is a cornerstone of perception and cognition. Computationally, categorization amounts to applying decision boundaries in the space of stimulus features. We designed a visual categorization task in which optimal performance requires observers to incorporate trial-to-trial knowledge of the level of sensory uncertainty when setting their decision boundaries. We found that humans and monkeys did adjust their decision boundaries from trial to trial as the level of sensory noise varied, with some subjects performing near optimally. We constructed a neural network that implements uncertainty-based, near-optimal adjustment of decision boundaries. Divisive normalization emerges automatically as a key neural operation in this network. Our results offer an integrated computational and mechanistic framework for categorization under uncertainty.

Original languageEnglish (US)
Pages (from-to)20332-20337
Number of pages6
JournalProceedings of the National Academy of Sciences of the United States of America
Volume110
Issue number50
DOIs
StatePublished - Dec 10 2013

Fingerprint

Neural networks
Uncertainty

Keywords

  • Bayesian inference
  • Decision-making
  • Optimality
  • Vision

ASJC Scopus subject areas

  • General

Cite this

Trial-to-trial, uncertainty-based adjustment of decision boundaries in visual categorization. / Qamar, Ahmad T.; Cotton, R. James; George, Ryan G.; Beck, Jeffrey M.; Prezhdo, Eugenia; Laudano, Allison; Tolias, Andreas S.; Ma, Wei Ji.

In: Proceedings of the National Academy of Sciences of the United States of America, Vol. 110, No. 50, 10.12.2013, p. 20332-20337.

Research output: Contribution to journalArticle

Qamar, Ahmad T. ; Cotton, R. James ; George, Ryan G. ; Beck, Jeffrey M. ; Prezhdo, Eugenia ; Laudano, Allison ; Tolias, Andreas S. ; Ma, Wei Ji. / Trial-to-trial, uncertainty-based adjustment of decision boundaries in visual categorization. In: Proceedings of the National Academy of Sciences of the United States of America. 2013 ; Vol. 110, No. 50. pp. 20332-20337.
@article{9cbaa2ef2eae491bb01b8fe1c3660f52,
title = "Trial-to-trial, uncertainty-based adjustment of decision boundaries in visual categorization",
abstract = "Categorization is a cornerstone of perception and cognition. Computationally, categorization amounts to applying decision boundaries in the space of stimulus features. We designed a visual categorization task in which optimal performance requires observers to incorporate trial-to-trial knowledge of the level of sensory uncertainty when setting their decision boundaries. We found that humans and monkeys did adjust their decision boundaries from trial to trial as the level of sensory noise varied, with some subjects performing near optimally. We constructed a neural network that implements uncertainty-based, near-optimal adjustment of decision boundaries. Divisive normalization emerges automatically as a key neural operation in this network. Our results offer an integrated computational and mechanistic framework for categorization under uncertainty.",
keywords = "Bayesian inference, Decision-making, Optimality, Vision",
author = "Qamar, {Ahmad T.} and Cotton, {R. James} and George, {Ryan G.} and Beck, {Jeffrey M.} and Eugenia Prezhdo and Allison Laudano and Tolias, {Andreas S.} and Ma, {Wei Ji}",
year = "2013",
month = "12",
day = "10",
doi = "10.1073/pnas.1219756110",
language = "English (US)",
volume = "110",
pages = "20332--20337",
journal = "Proceedings of the National Academy of Sciences of the United States of America",
issn = "0027-8424",
number = "50",

}

TY - JOUR

T1 - Trial-to-trial, uncertainty-based adjustment of decision boundaries in visual categorization

AU - Qamar, Ahmad T.

AU - Cotton, R. James

AU - George, Ryan G.

AU - Beck, Jeffrey M.

AU - Prezhdo, Eugenia

AU - Laudano, Allison

AU - Tolias, Andreas S.

AU - Ma, Wei Ji

PY - 2013/12/10

Y1 - 2013/12/10

N2 - Categorization is a cornerstone of perception and cognition. Computationally, categorization amounts to applying decision boundaries in the space of stimulus features. We designed a visual categorization task in which optimal performance requires observers to incorporate trial-to-trial knowledge of the level of sensory uncertainty when setting their decision boundaries. We found that humans and monkeys did adjust their decision boundaries from trial to trial as the level of sensory noise varied, with some subjects performing near optimally. We constructed a neural network that implements uncertainty-based, near-optimal adjustment of decision boundaries. Divisive normalization emerges automatically as a key neural operation in this network. Our results offer an integrated computational and mechanistic framework for categorization under uncertainty.

AB - Categorization is a cornerstone of perception and cognition. Computationally, categorization amounts to applying decision boundaries in the space of stimulus features. We designed a visual categorization task in which optimal performance requires observers to incorporate trial-to-trial knowledge of the level of sensory uncertainty when setting their decision boundaries. We found that humans and monkeys did adjust their decision boundaries from trial to trial as the level of sensory noise varied, with some subjects performing near optimally. We constructed a neural network that implements uncertainty-based, near-optimal adjustment of decision boundaries. Divisive normalization emerges automatically as a key neural operation in this network. Our results offer an integrated computational and mechanistic framework for categorization under uncertainty.

KW - Bayesian inference

KW - Decision-making

KW - Optimality

KW - Vision

UR - http://www.scopus.com/inward/record.url?scp=84890287562&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84890287562&partnerID=8YFLogxK

U2 - 10.1073/pnas.1219756110

DO - 10.1073/pnas.1219756110

M3 - Article

VL - 110

SP - 20332

EP - 20337

JO - Proceedings of the National Academy of Sciences of the United States of America

JF - Proceedings of the National Academy of Sciences of the United States of America

SN - 0027-8424

IS - 50

ER -