QUANTAL RELEASE OF TRANSMITTER AT THE FROG CYTONEURAL JUNCTION DURING SUSTAINED ACTIVATION

The mechanism of quantal release at the posterior canal cytoneural junction was investigated by studying the size and interval distributions of mEPSPs at rest and during excitatory mechanical stimulation (acceleration: 2-7 deg.s-2) or electrical activation (10-40 Hz) of excitatory efferent system of the canal. Resting rates of mEPSPs are usually high (>100/s) leading to extensive overlapping of individual events. When a large number of well defined unitary events (25-100) could be detected they were aligned and the averaged waveform was fitted by a modified gamma distribution function. At high mEPSP rates noise analysis was exploited. According to the theory, the autocotrelation of a signal generated by the random summation of reasonably homogeneous elementary events has the same shape as the autocorrelation of the underlying single event. Previous investigations indicated that the mEPSP waveform is minimum phase; therefore the autoregressive (minimum phase) fit to the autocorrelation yields the analytical approximation to the event waveform. The fitted mEPSP waveform was used to construct a Wiener filter to deconvolve the junctional recordings. After filtering tracings with overlapping potentials display a pattern of well resolved peaky events where the location of each peak indicates the occurrence of a mEPSP and its size is proportional to the original event amplitude. When overlapping was negligible, waveforms and size histograms constructed from the original recordings or by the autocorrelation + Wiener filtering procedure were in good agreement. When overlapping was extensive, the Wiener filter eliminated the apparent excess of oversized events. All size histograms were well fitted by lognormal distribution functions. Time interval histograms of the resting mEPSP were consistently fitted by a single exponential, suggesting random occurrence; this was not the case for interval histograms during stimulation. These results are in agreement with previous observations under reduced transmitter release rate and demonstrate that: 1) most apparently oversized events are due to the random summation of independent mEPSPs; 2) though release is activated by rotation or by changes in hair cell conductance operated by the efferent transmitter, the underlaying quantal mechanism is unaffected.