Encoding Multisensory Information in Modular Neural Networks

He Wang, Wen Hao Zhang, K. Y.Michael Wong, Si Wu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations


The brain is capable of integrating information in multiple sensory channels in a Bayesian optimal way. Based on a decentralized network model inspired by electrophysiological recordings, we consider the structural pre-requisites for optimal multisensory integration. In this architecture, same-channel feedforward and recurrent links encode the unisensory likelihoods, whereas reciprocal couplings connecting the different modules are shaped by the correlation in the joint prior probabilities. Moreover, the statistical relationship between the difference in the optimal network structures and the difference in the priors and the likelihoods clearly shows that the network can encode multisensory information in a distributed manner. Our results generate testable predictions for future experiments and are likely to be applicable to other artificial systems.

Original languageEnglish (US)
Title of host publicationNeural Information Processing - 24th International Conference, ICONIP 2017, Proceedings
EditorsDerong Liu, Shengli Xie, Yuanqing Li, El-Sayed M. El-Alfy, Dongbin Zhao
PublisherSpringer Verlag
Number of pages8
ISBN (Print)9783319700922
StatePublished - 2017
Externally publishedYes
Event24th International Conference on Neural Information Processing, ICONIP 2017 - Guangzhou, China
Duration: Nov 14 2017Nov 18 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10637 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference24th International Conference on Neural Information Processing, ICONIP 2017


  • Multisensory Bayesian inference
  • Recurrent neural networks

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'Encoding Multisensory Information in Modular Neural Networks'. Together they form a unique fingerprint.

Cite this