Cortical Plasticity as a Mechanism for Storing Bayesian Priors in Sensory Perception
Both the naïve model AI (a and c) and 7-kHz-over-represented AI (b and d) were examined with (c and d) and without (a and b) elevated baseline activity. In addition, standard deviation of the decoded frequencies (red) was used to measure the output variability. When baseline activity was elevated in the 7-kHz-over-represented AI, the decoded frequencies show shifts characteristic of Bayesian prior bias (d). The pink line shows the slope of the input-output curve at the over-represented frequency. The slope is a measure of the prior bias.