Recovery of auditory-nerve-fiber spike amplitude under natural excitation conditions

Knowledge of the refractory properties of auditory-nerve fibers (ANFs) is required for understanding the transduction of the graded membrane potential of the receptor cells into spike trains. The refractory properties inferred when ANFs are excited by electrical stimulation might differ from those present when ANFs are excited naturally by transmitter release from receptor cells. As a proxy for the latter, we investigated the recovery of spike amplitude with time since the previous spike in long extracellular recordings of the activity of individual ANFs from anesthetized Mongolian gerbils. Voltage traces were filtered minimally to avoid distortions of spike amplitude and timing. The amplitude of each spike was defined as the difference between its peak voltage and an extrapolated instantaneous reference voltage at the time of the peak. Spike amplitude was normalized to that of the previous spike to exclude effects of long-term changes in recording conditions. To ensure that the amplitude of the first spike in each pair was fully recovered, each spike pair was used only when preceded by an interspike interval of at least 5 ms. We find that the recovery of spike amplitude is well described by a short dead time followed by a double-exponential recovery function. Total recovery times were short (median: 0.85 ms; interquartile range: 0.74-1.00 ms) and independent of the ANF's characteristic frequency and spontaneous rate, but they increased weakly with increasing mean rate. We emphasize the differences between the recovery of spike amplitude, the recovery of spike probability from postsynaptic refractoriness, and the recovery of spike probability as reflected in the hazard-rate function. Our findings are inconsistent with the long refractory periods assumed in some models, but are consistent with the brief refractoriness assumed in the synapse model of Peterson and Heil (2018), which reproduces the stochastic properties of stationary spontaneous and sound-driven ANF spike trains.