THE FREQUENCY DEPENDENCE OF TRANSMITTER RELEASE, IN RESPONSE TO SINUSOIDAL ROTATION, AT THE CYTONEURAL JUNCTION IN FROG LABYRINTH

Classically, the response of the semicircular canal to rotational stimulation has been studied in terms of the spike firing pattern of the primary afferents. However this is the final result of a series of signal processing steps. These involve the hydrodynamic properties of the cupula-endolymph system, the processes of mechano-electrical transduction, regulation of membrane potential and transmitter release by the hair cell, and the properties of the spike encoder which translates the time varying postsynaptic potential into action potentials. The regulation of transmitter release by the hair cell has been the subject of few studies. We have previousiy demonstrated that quantitative measurements of quantal release (rates of occurrence of excitatory postsynaptic potentials, EPSPs) can be obtained by means of noise analysis during intense mechanical stimulation of the canal. Transmitter release was studied at the frog posterior canal cytoneural junction, in response to sinusoidal rotation with varying intensity at fixed frequency (0.1 Hz), and it was concluded that asymmetry, adaptation and at least part of the phase lead which are generally observed in the firing pattern of the primary afferents are of presynaptic origin, as they are observed in the pattern of transmitter release as well. Conversely, rectification (silencing of spikes during inhibitory rotation) and a possible further phase lead appeared to arise at the encoder. In this study the characterization of presynaptic activity has been completed by measuring the response to stimuli of relatively constant intensity (peak acceleration: 15-18 deg-s-2) but at different frequencies (0.02 to 0.5 Hz) and to rotations at varying frequency (0.02-0.5 Hz) and acceleration (2.5-63 deg-s-2). Thus, the frequency response of transmitter release at the cytoneural junction has been defined. The waveform of EPSPs was evaluated by power spectral analysis; EPSP amplitude and rate of occurrence were determined by means of a procedure of fluctuation analysis devised to continuously monitor these parameters in a rapidly changing system, with a time resolution such that rotation frequencies up to 1Hz could be studied. Asymmetry in the response was generally observed at all rotation frequencies, i.e. the increase in EPSP rate in response to excitatory acceleration was systematically larger than the decrease produced by the corresponding inhibitory acceleration. At all frequencies of rotation the time course of EPSP rate of occurrence was well described by analytical *asymmetrical" sinusoid. For each unit and at each frequency of sinusoidal rotation it was possible to compute a “gain" of the response (peak-to-peak swing in EPSP rate per unit peak acceleration), a *phase" (lead with respect to angular acceleration or velocity) and an *asymmetry factor" (difference between excitatory and inhibitory responses, divided by peak-to-peak swing). In the experiments with fixed peak acceleration the gain-vs.-frequency relations were well described by one-pole low-pass fílter functions. The estimated corner frequencies ranged 0.04-0.24 Hz in the different units, phase leads being generally consistent with the functions fit to gain data. However, the confidence limits for these estimates are compatible with the idea that the frequency dependence may be common to all units and determined by the hydrodynamic properties of the cupula-endolymph system. Asymmetry in the response did not display any definite frequency-dependence. The same behaviour was observed in those units where peak accelerations were changed up to 25-fold, as expected from the linearity of the stimulus-response relations in the range of stimulation intensities here explored. From the frequency responses the static gain (zero frequency extrapolation) of each unit was defined. A decrease in the response to successive cycles of rotation (adaptation) was observed when the EPSP rate during the first cycle exceeded in each unit a value proportional to its own static gain. No well defined patterns or dominant frequencies of oscillation (tuning) could be identified in the fluctuations of EPSP rate, either at rest or in the residuals from fits during rotation. This study constitutes the first attempt at estimating the frequency response of neurotransmitter release by the hair cell. The results here reported indicate that most of the properties of spike firing at the cytoneural junction directly follow from the properties of neurotransmitter release by the presynaptic cell.