Neurotransmitter release statistics: Moment estimates for inhomogeneous Bernoulli trials

Summary Evoked release of quanta of neurotransmitter is generally treated as a set of homogeneous, stationary Bernoulli trials, hence governed by the binomial distribution. Relaxing the assumptions of uniformity and stationarity leads to a more realistic physiological model of transmitter release but also introduces systematic biases in the moment estimates of the binomial parameters. We derive probability generating functions for quantal release and expressions for the moment estimates of ¯n and ¯p for a generalized model that incorporates temporal variation and nonuniformity in individual release probabilities and in numbers of release sites.