Bayesian statistics: relevant for the brain?

Abstract

Analyzing data from experiments involves variables that we neuroscientists are uncertain about. Efficiently calculating with such variables usually requires Bayesian statistics. As it is crucial when analyzing complex data, it seems natural that the brain would “use” such statistics to analyze data from the world. And indeed, recent studies in the areas of perception, action, and cognition suggest that Bayesian behavior is widespread, in many modalities and species. Consequently, many models have suggested that the brain is built on simple Bayesian principles. While the brain’s code is probably not actually simple, I believe that Bayesian principles will facilitate the construction of faithful models of the brain.

Bayesian data analysis for newcomers

“…In the psychological sciences, Bayesian models have been used extensively not for data analysis but for models of cognition and brain function (e.g., Chater & Oaksford, 2008;Chater, Tenenbaum, & Yuille, 2006;Griffiths, Kemp, & Tenenbaum, 2008;Gilet, Diard, & Bessière, 2011;Jacobs & Kruschke, 2010;Kording, 2014;Kruschke, 2008;Perfors, Tenenbaum, Griffiths, & Xu, 2011). Many newcomers to Bayesian data analysis may have previous exposure to Bayesian models of mind and brain.…”

An Analysis of a Ring Attractor Model for Cue Integration

“…Bayes' theorem (1) provides a mathematical framework describing the optimal way in which information from different sources should be combined, and some studies argue that animals have Bayesian brains [2,[5][6][7]. According to Bayes' Rule, the posterior probability P (x true |x cue ) (the probability of event x will happen when the cue about x is sensed) is proportional to the product of the prior probability P (x true ) (the probability of event x happening based on prior knowledge) and the likelihood function P (x cue |x true ) (the probability of the cue when x truly happened, which represents the reliability of this cue).…”

The spatio‐temporal profile of multisensory integration

“…Importantly, these behavioural benefits appear to be more pronounced if unisensory inputs are weak, that is less reliable (Nath & Beauchamp, 2011;Rohe & Noppeney, 2015), in accordance with the neural principle of inverse effectiveness (Meredith & Stein, 1983;Stein & Meredith, 1994). On a computational level, the weighting of multisensory inputs by their reliability was first conceptually described as modality appropriateness (Welch & Warren, 1986) and was later formalised mathematically by maximum likelihood estimation (Ernst & Banks, 2002) and Bayesian causal modelling (Kording, 2014). These methods can successfully predict a range of multisensory phenomena by the weighting of sensory inputs.…”