Mathematical modelling of reaction latency. Part II: A model of escape reaction and escape learning.

The results of fitting the linear-dynamic-stochastic model to the latency data from escape conditioning experiment are presented. Two processes can be distinguished: (i) an increase in parameters of the dynamic part of the model (called the static gain and the dynamic gain) which results in a decrease in the reaction latency during learning, (ii) an increase in parameter of the decision part of the model (mean value of the threshold), which results simultaneously in an increase in the reaction latency and a decrease in the inter-trial responses probability. The reaction latency decreases during the learning because the first phenomenon prevails over the second one, but the latter is the only reason of inter-trial responses decrease during learning. Analysis of the results of surgical treatment on the learning process is also presented.