Is loudness simply proportional to the auditory nerve spike count?

It is often asserted that the physiological correlate of loudness is the simple sum of the spike activity produced by all neurons in the auditory nerve (the auditory nerve spike count). We will refer to this hypothesis as the spike count hypothesis. The spike count hypothesis has been tested in the past using models of the auditory periphery and in almost all cases, the hypothesis has been supported. Our new technique for recording a compound potential from the chinchilla auditory nerve, the perstimulus compound action potential (PCAP), makes possible the measurement of the growth of the auditory nerve spike count, thus providing data that can be used to test the spike count hypothesis empirically. It was observed that the growth of the auditory nerve spike count in response to a 1-kHz pure tone (in dB/dB) is 33% shallower than the growth of loudness for a 1-kHz tone, and this discrepancy increases to 66% for an 8-kHz tone. In addition, ''equal-count'' contours were constructed in a manner analogous to equal-loudness contours. It was found that as reference intensity increases, equal-count contours become sharply curved upward at high frequencies whereas equal-loudness contours become increasingly flat. These differences are unlikely to be the result of the cross-species comparison, since the discrepancies are mostly attributable to the skewed pattern of spread of excitation along the basilar membrane, a property shared by humans and chinchillas. Therefore, we conclude that the simple sum of the spike activity in the auditory nerve cannot be the physiological correlate of loudness. (C) 1997 Acoustical Society of America.