Implementation of a Bayesian Filter in a Photoreceptor Cell

Overview & Synopsis

In this study we consider how probability distributions calculated in probabilistic cognitive models can be represented and processed in the brain. More exactly we show that a photoreceptor cell can compute a simple Bayesian inference in a binary Hidden Markov Model (HMM), based on the underlying biochemical interactions of this single cell. We derive, under steady-state conditions, a formal equivalence between the probabilistic model and the molecular mechanisms, and show that the equivalence can be extended to the dynamic case. From the photoreceptor example we see that biochemical interactions can represent probability distributions and implement basic probabilistic reasoning.