Biomedical Engineering Reference
In-Depth Information
Chapter 4
Computational Analysis of Behavioural
and Neural Data Through Bayesian Statistical
Modelling
Raymond Wai Mun Chan and Fabrizio Gabbiani
Abstract
Computational analysis of behavioural and neural data is nowadays an
essential part of neuroethology, allowing an ever deeper understanding of how natural
behaviour and neural activity are interrelated at the molecular, cellular and network
level. The range of computational techniques applied in neuroethological research is
currently so broad as to preclude an exhaustive survey in a succinct chapter. Here, we
focus on a speciic approach termed Bayesian statistical modelling that has proven to
be a powerful method for relating neural activity to natural behavioural performance.
As we illustrate in a speciic example, this approach naturally dovetails with classic
neural coding concepts such as population vector codes. It is also lexible enough to
be applicable to a broad range of neuroethological questions.
Keywords
Barn owl • Bayesian models • Interaural time difference • Maximum
likelihood • Neural correlations • Population codes • Population vector • Sound
localization
