Learning flexible sensori-motor mappings in a complex network

Eleni Vasilaki, Stefano Fusi, Xiao-Jing Wang, Walter Senn

Research output: Contribution to journalArticle

Abstract

Given the complex structure of the brain, how can synaptic plasticity explain the learning and forgetting of associations when these are continuously changing? We address this question by studying different reinforcement learning rules in a multilayer network in order to reproduce monkey behavior in a visuomotor association task. Our model can only reproduce the learning performance of the monkey if the synaptic modifications depend on the pre- and postsynaptic activity, and if the intrinsic level of stochasticity is low. This favored learning rule is based on reward modulated Hebbian synaptic plasticity and shows the interesting feature that the learning performance does not substantially degrade when adding layers to the network, even for a complex problem.

Original languageEnglish (US)
Pages (from-to)147-158
Number of pages12
JournalBiological Cybernetics
Volume100
Issue number2
DOIs
StatePublished - Feb 2009

Fingerprint

Complex networks
Plasticity
Learning
Neuronal Plasticity
Reinforcement learning
Haplorhini
Brain
Multilayers
Association Learning
Reward

Keywords

  • Hebbian
  • Learning
  • Multilayer
  • Reward-modulated
  • Visuomotor task

ASJC Scopus subject areas

  • Biotechnology
  • Computer Science(all)

Cite this

Learning flexible sensori-motor mappings in a complex network. / Vasilaki, Eleni; Fusi, Stefano; Wang, Xiao-Jing; Senn, Walter.

In: Biological Cybernetics, Vol. 100, No. 2, 02.2009, p. 147-158.

Research output: Contribution to journalArticle

Vasilaki, Eleni ; Fusi, Stefano ; Wang, Xiao-Jing ; Senn, Walter. / Learning flexible sensori-motor mappings in a complex network. In: Biological Cybernetics. 2009 ; Vol. 100, No. 2. pp. 147-158.
@article{c15be4b74bc74a50a7ff38e769e2885d,
title = "Learning flexible sensori-motor mappings in a complex network",
abstract = "Given the complex structure of the brain, how can synaptic plasticity explain the learning and forgetting of associations when these are continuously changing? We address this question by studying different reinforcement learning rules in a multilayer network in order to reproduce monkey behavior in a visuomotor association task. Our model can only reproduce the learning performance of the monkey if the synaptic modifications depend on the pre- and postsynaptic activity, and if the intrinsic level of stochasticity is low. This favored learning rule is based on reward modulated Hebbian synaptic plasticity and shows the interesting feature that the learning performance does not substantially degrade when adding layers to the network, even for a complex problem.",
keywords = "Hebbian, Learning, Multilayer, Reward-modulated, Visuomotor task",
author = "Eleni Vasilaki and Stefano Fusi and Xiao-Jing Wang and Walter Senn",
year = "2009",
month = "2",
doi = "10.1007/s00422-008-0288-z",
language = "English (US)",
volume = "100",
pages = "147--158",
journal = "Biological Cybernetics",
issn = "0340-1200",
publisher = "Springer Verlag",
number = "2",

}

TY - JOUR

T1 - Learning flexible sensori-motor mappings in a complex network

AU - Vasilaki, Eleni

AU - Fusi, Stefano

AU - Wang, Xiao-Jing

AU - Senn, Walter

PY - 2009/2

Y1 - 2009/2

N2 - Given the complex structure of the brain, how can synaptic plasticity explain the learning and forgetting of associations when these are continuously changing? We address this question by studying different reinforcement learning rules in a multilayer network in order to reproduce monkey behavior in a visuomotor association task. Our model can only reproduce the learning performance of the monkey if the synaptic modifications depend on the pre- and postsynaptic activity, and if the intrinsic level of stochasticity is low. This favored learning rule is based on reward modulated Hebbian synaptic plasticity and shows the interesting feature that the learning performance does not substantially degrade when adding layers to the network, even for a complex problem.

AB - Given the complex structure of the brain, how can synaptic plasticity explain the learning and forgetting of associations when these are continuously changing? We address this question by studying different reinforcement learning rules in a multilayer network in order to reproduce monkey behavior in a visuomotor association task. Our model can only reproduce the learning performance of the monkey if the synaptic modifications depend on the pre- and postsynaptic activity, and if the intrinsic level of stochasticity is low. This favored learning rule is based on reward modulated Hebbian synaptic plasticity and shows the interesting feature that the learning performance does not substantially degrade when adding layers to the network, even for a complex problem.

KW - Hebbian

KW - Learning

KW - Multilayer

KW - Reward-modulated

KW - Visuomotor task

UR - http://www.scopus.com/inward/record.url?scp=63149146163&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=63149146163&partnerID=8YFLogxK

U2 - 10.1007/s00422-008-0288-z

DO - 10.1007/s00422-008-0288-z

M3 - Article

VL - 100

SP - 147

EP - 158

JO - Biological Cybernetics

JF - Biological Cybernetics

SN - 0340-1200

IS - 2

ER -