Origin of latency shift of cochlear nerve potentials with sound intensity

Compound action potentials were recorded from the round window in anesthetized rats in response to tonebursts and to continuous tones. The responses to tonebursts were compared to cross-correlograms of the responses to continuous tones that were amplitude modulated by pseudorandom noise. The cross-correlograms were similar in shape to the responses to tonebursts, but the latencies of the two cross-correlogram peaks decreased less when the sound intensity was increased from near threshold values than did the latencies of the peaks in the responses to tonebursts. When an unmodulated tone was added to these stimuli, the latencies of peaks in the response to tonebursts increased as the intensity of the unmodulated tone was increased. However, the effect of an unmodulated tone on the latencies of the peaks in the cross-correlograms was more complex: when the frequency of the unmodulated tone was below that of the modulated tone, there was a decrease in the latency of the peaks in cross-correlograms. This decrease in latency was a function of the intensity of the unmodulated tone, and was similar to the decrease in latency seen when the intensity of a single modulated tone is increased. These results are interpreted as supporting the findings of earlier research, which showed that the cross-correlogram of the unit response to amplitude-modulated tones has a latency that is nearly independent of the stimulus intensity. In addition, the data indicate that the decrease in the latency of the peaks in the cross-correlograms of the gross response from the round window is a result of nonlinearity of the cochlear frequency analyzer and is not directly related to neural excitation in the cochlea.