Effects of variance in mini amplitude on stimulus-evoked release: a comparison of two models.

The strength of synaptic connections between two neurons is characterized by the number of release sites (N) on the presynaptic cell, the probability (p) of transmitter release at those sites in response to a stimulus, and the average size (A) of the postsynaptic response from each site. Quantal analysis can determine N, p, and A, but the large variance in the amplitudes of minis at central synapses is predicted to obscure quantal peaks and render quantal analysis unusable. Recently it has been suggested that the variance in mini amplitude is generated by differences between release sites, rather than by quantum-to-quantum fluctuations at identical sites, and that this form of variance in mini amplitude reduces the amount of variance expected in quantal peaks. Using simulations, we examine the possibility of resolving quantal peaks assuming either form of variance in mini amplitude. We find that individual quantal peaks are resolvable in neither case, provided that the uniquantal distribution is similar to the mini distribution. Because this lack of resolution compromises the utility of quantal analysis, we develop a general description that can solve N and p, given the statistical parameters of the mini distribution and the evoked distribution. We find that this description is relatively insensitive to the source of variance in mini amplitude.