Nonuniform release probabilities underlie quantal synaptic transmission at a mammalian excitatory central synapse.

1. Excitatory postsynaptic potentials (EPSPs) evoked by impulses in single group I muscle afferents were recorded in dorsal spinocerebellar tract (DSCT) neurons in the spinal cords of anesthetized cats. Fluctuations in the amplitude of these single-fiber EPSPs were determined from measurements of EPSP peak amplitude and contaminating noise (800-4600 trials). 2. In a previous study at this connection, we found that these single-fiber EPSPs fluctuated in amplitude between approximately equal, or quantal, increments. However, these quantal fluctuations could not be described by simple binomial statistics (39). In the present study we have applied further analysis procedures to the same single-fiber EPSPs to formulate a more appropriate probabilistic model of transmission at this connection. 3. In the first stage we have demonstrated that each single-fiber EPSP is composed of the sum of a number (3-30) of uniform quantal events, and that there is extremely little variability in the amplitude of the single quantal event. 4. In a further procedure, we have demonstrated that these quantal fluctuations can be described by a compound binomial model in which each underlying quantal event is associated with a particular, but independent, release probability. The results of this analysis indicate that the probability of transmitter release varies considerably between release sites at this connection. (The use of such a compound binomial model reemphasized previous warnings concerning the interpretation of the results of all statistical models of quantal release. Problems regarding the non-unique nature of N, the total population of quantal events, and other such difficulties are discussed.) 5. A model of transmission at this connection is proposed, in which there are a number of "active" release sites, exhibiting generally high release probabilities, and a number of "reserve" release sites, with zero, or close to zero, release probability. The physiological consequences of such a scheme are discussed.