Stochastic modeling of facilitated neurosecretion.

Two models of neurosecretion were evaluated in terms of their ability to predict the dependency of quantal content (m) on the frequency of repetitive stimulation of a lobster motoneuron. First, the hypothesis that neurosecretion is limited by a fixed number of release sites was tested by the fit of the distribution of m by uniform and nonuniform binomial statistics. The obtained release probabilities suggest that frequency facilitation can be due to activation of a group of sites with high release probabilities. However, the fit obtained using this model is not statistically significant due to a large number of fitting parameters. Second, the hypothesis that neurosecretion is limited by the rates of exchange between the releasable pool and the total store of quanta and that each stimulus enhances quantal mobilization was tested. Monte Carlo simulation was carried out in accordance with this model and reproduced the observed distribution of m with very few fitting parameters and therefore with a high level of significance (>0.1). This result demonstrates that mobilization of extra vesicles with each stimulus is a mechanism that allows a very accurate and parsimonious quantitative description of frequency facilitation.