Bayesian inference is a powerful methodology but can be over-extended. Above is an example of over-extending the Bayes framework to neuroscience. Given its power, Bayes inference can easily account for behavioral phenomena, but it does not provide a plausible neural mechanism. There are no Bayesian neurons or areas in the brain.
Starting with the underlying architecture of the brain (i.e., neurons, synapses, and neurotransmitters) puts neuroscience on a path to develop more robust models. Models built from the underlying neural architecture of the brain would be more complex than the simple elegance of Bayesian models but provide the opportunity for deeper insights into real brain processes.
Starting with the underlying architecture of the brain (i.e., neurons, synapses, and neurotransmitters) puts neuroscience on a path to develop more robust models. Models built from the underlying neural architecture of the brain would be more complex than the simple elegance of Bayesian models but provide the opportunity for deeper insights into real brain processes.
No comments:
Post a Comment