A dual role for prediction error in associative learning

Hanneke E M Den Ouden, Karl J. Friston, Nathaniel D. Daw, Anthony R. McIntosh, Klaas E. Stephan

Research output: Contribution to journalArticle

Abstract

Confronted with a rich sensory environment, the brain must learn statistical regularities across sensory domains to construct causal models of the world. Here, we used functional magnetic resonance imaging and dynamic causal modeling (DCM) to furnish neurophysiological evidence that statistical associations are learnt, even when task-irrelevant. Subjects performed an audio-visual target-detection task while being exposed to distractor stimuli. Unknown to them, auditory distractors predicted the presence or absence of subsequent visual distractors. We modeled incidental learning of these associations using a Rescorla-Wagner (RW) model. Activity in primary visual cortex and putamen reflected learning-dependent surprise: these areas responded progressively more to unpredicted, and progressively less to predicted visual stimuli. Critically, this prediction-error response was observed even when the absence of a visual stimulus was surprising. We investigated the underlying mechanism by embedding the RW model into a DCM to show that auditory to visual connectivity changed significantly over time as a function of prediction error. Thus, consistent with predictive coding models of perception, associative learning is mediated by prediction-error dependent changes in connectivity. These results posit a dual role for prediction-error in encoding surprise and driving associative plasticity.

Original languageEnglish (US)
Pages (from-to)1175-1185
Number of pages11
JournalCerebral Cortex
Volume19
Issue number5
DOIs
StatePublished - May 2009

Fingerprint

Association Learning
Learning
Putamen
Visual Cortex
Magnetic Resonance Imaging
Brain

Keywords

  • Associative learning
  • Cross-modal
  • Dynamic causal modeling
  • Effective connectivity
  • FMRI
  • Rescorla-Wagner model

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Cellular and Molecular Neuroscience
  • Medicine(all)

Cite this

Den Ouden, H. E. M., Friston, K. J., Daw, N. D., McIntosh, A. R., & Stephan, K. E. (2009). A dual role for prediction error in associative learning. Cerebral Cortex, 19(5), 1175-1185. https://doi.org/10.1093/cercor/bhn161

A dual role for prediction error in associative learning. / Den Ouden, Hanneke E M; Friston, Karl J.; Daw, Nathaniel D.; McIntosh, Anthony R.; Stephan, Klaas E.

In: Cerebral Cortex, Vol. 19, No. 5, 05.2009, p. 1175-1185.

Research output: Contribution to journalArticle

Den Ouden, HEM, Friston, KJ, Daw, ND, McIntosh, AR & Stephan, KE 2009, 'A dual role for prediction error in associative learning', Cerebral Cortex, vol. 19, no. 5, pp. 1175-1185. https://doi.org/10.1093/cercor/bhn161
Den Ouden HEM, Friston KJ, Daw ND, McIntosh AR, Stephan KE. A dual role for prediction error in associative learning. Cerebral Cortex. 2009 May;19(5):1175-1185. https://doi.org/10.1093/cercor/bhn161
Den Ouden, Hanneke E M ; Friston, Karl J. ; Daw, Nathaniel D. ; McIntosh, Anthony R. ; Stephan, Klaas E. / A dual role for prediction error in associative learning. In: Cerebral Cortex. 2009 ; Vol. 19, No. 5. pp. 1175-1185.
@article{44559c0a3c2f4691b45f7b88f753cb78,
title = "A dual role for prediction error in associative learning",
abstract = "Confronted with a rich sensory environment, the brain must learn statistical regularities across sensory domains to construct causal models of the world. Here, we used functional magnetic resonance imaging and dynamic causal modeling (DCM) to furnish neurophysiological evidence that statistical associations are learnt, even when task-irrelevant. Subjects performed an audio-visual target-detection task while being exposed to distractor stimuli. Unknown to them, auditory distractors predicted the presence or absence of subsequent visual distractors. We modeled incidental learning of these associations using a Rescorla-Wagner (RW) model. Activity in primary visual cortex and putamen reflected learning-dependent surprise: these areas responded progressively more to unpredicted, and progressively less to predicted visual stimuli. Critically, this prediction-error response was observed even when the absence of a visual stimulus was surprising. We investigated the underlying mechanism by embedding the RW model into a DCM to show that auditory to visual connectivity changed significantly over time as a function of prediction error. Thus, consistent with predictive coding models of perception, associative learning is mediated by prediction-error dependent changes in connectivity. These results posit a dual role for prediction-error in encoding surprise and driving associative plasticity.",
keywords = "Associative learning, Cross-modal, Dynamic causal modeling, Effective connectivity, FMRI, Rescorla-Wagner model",
author = "{Den Ouden}, {Hanneke E M} and Friston, {Karl J.} and Daw, {Nathaniel D.} and McIntosh, {Anthony R.} and Stephan, {Klaas E.}",
year = "2009",
month = "5",
doi = "10.1093/cercor/bhn161",
language = "English (US)",
volume = "19",
pages = "1175--1185",
journal = "Cerebral Cortex",
issn = "1047-3211",
publisher = "Oxford University Press",
number = "5",

}

TY - JOUR

T1 - A dual role for prediction error in associative learning

AU - Den Ouden, Hanneke E M

AU - Friston, Karl J.

AU - Daw, Nathaniel D.

AU - McIntosh, Anthony R.

AU - Stephan, Klaas E.

PY - 2009/5

Y1 - 2009/5

N2 - Confronted with a rich sensory environment, the brain must learn statistical regularities across sensory domains to construct causal models of the world. Here, we used functional magnetic resonance imaging and dynamic causal modeling (DCM) to furnish neurophysiological evidence that statistical associations are learnt, even when task-irrelevant. Subjects performed an audio-visual target-detection task while being exposed to distractor stimuli. Unknown to them, auditory distractors predicted the presence or absence of subsequent visual distractors. We modeled incidental learning of these associations using a Rescorla-Wagner (RW) model. Activity in primary visual cortex and putamen reflected learning-dependent surprise: these areas responded progressively more to unpredicted, and progressively less to predicted visual stimuli. Critically, this prediction-error response was observed even when the absence of a visual stimulus was surprising. We investigated the underlying mechanism by embedding the RW model into a DCM to show that auditory to visual connectivity changed significantly over time as a function of prediction error. Thus, consistent with predictive coding models of perception, associative learning is mediated by prediction-error dependent changes in connectivity. These results posit a dual role for prediction-error in encoding surprise and driving associative plasticity.

AB - Confronted with a rich sensory environment, the brain must learn statistical regularities across sensory domains to construct causal models of the world. Here, we used functional magnetic resonance imaging and dynamic causal modeling (DCM) to furnish neurophysiological evidence that statistical associations are learnt, even when task-irrelevant. Subjects performed an audio-visual target-detection task while being exposed to distractor stimuli. Unknown to them, auditory distractors predicted the presence or absence of subsequent visual distractors. We modeled incidental learning of these associations using a Rescorla-Wagner (RW) model. Activity in primary visual cortex and putamen reflected learning-dependent surprise: these areas responded progressively more to unpredicted, and progressively less to predicted visual stimuli. Critically, this prediction-error response was observed even when the absence of a visual stimulus was surprising. We investigated the underlying mechanism by embedding the RW model into a DCM to show that auditory to visual connectivity changed significantly over time as a function of prediction error. Thus, consistent with predictive coding models of perception, associative learning is mediated by prediction-error dependent changes in connectivity. These results posit a dual role for prediction-error in encoding surprise and driving associative plasticity.

KW - Associative learning

KW - Cross-modal

KW - Dynamic causal modeling

KW - Effective connectivity

KW - FMRI

KW - Rescorla-Wagner model

UR - http://www.scopus.com/inward/record.url?scp=67649185192&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67649185192&partnerID=8YFLogxK

U2 - 10.1093/cercor/bhn161

DO - 10.1093/cercor/bhn161

M3 - Article

VL - 19

SP - 1175

EP - 1185

JO - Cerebral Cortex

JF - Cerebral Cortex

SN - 1047-3211

IS - 5

ER -