[Estimation of the excitation reduction following spike generation in the frog central auditory neurons]

A method for the estimation of the spike probability reduction caused by a preceding spike is proposed for central auditory neurons. The method is based upon a comparison of the normalized poststimulus time histogram with the normalized latency histogram and the expected probability function. For the solution of the superfluous system of linear equations, the least-squares fit was used. The result (temporal course of discharge history effect) was then approximated by a function that included several initial zero points (absolute refractory period). Three points were approximated by the error function (short relative refractory period) and all following points were equal to one. The only variable parameter was the number of zero points, the duration of refractoriness. This method was tested on a descriptive model of an auditory neuron in which each interspike interval consists of an absolute refractory period and a random-length waiting time to the next spike. The estimation of the model absolute refractory period by the proposed method was in reasonable coincidence with input data after several thousands of stimulus presentations. The same method was used for the estimation of dead-time in 87 single units of the midbrain auditory region of the frog. A great variability of refractory time (from less than 2 ms to more than 18 ms) was shown. The duration of refractoriness correlated with the latency period of neurons.