Auditory-nerve rate responses are inconsistent with common hypotheses for the neural correlates of loudness recruitment

A number of perceptual phenomena related to normal and impaired level coding can be accounted for by the degree of compression in the basilar-membrane (BM) magnitude response. However, the narrow dynamic ranges of auditory-nerve (AN) fibers complicate these arguments. Because the AN serves as an information bottleneck, an improved understanding of the neural coding of level may clarify some of the limitations of current hearing aids. Here three hypotheses for the neural correlate of loudness recruitment were evaluated based on AN responses from normal-hearing cats and from cats with a noise-induced hearing loss (NIHL). Auditory-nerve fiber rate-level functions for tones were analyzed to test the following hypotheses: Loudness recruitment results from steeper AN rate functions after impairment. This hypothesis was not supported; AN rate functions were not steeper than normal following NIHL, despite steeper estimated BM responses based on the AN data. Loudness is based on the total AN discharge count, and recruitment results from an abnormally rapid spread of excitation after impairment. Whereas abnormal spread of excitation can be observed, steeper growth of total AN rate is not seen over the range of sound levels where recruitment is observed in human listeners. Loudness of a narrowband stimulus is based on AN responses in a narrow BF region, and recruitment results from compression of the AN-fiber threshold distribution after impairment. This hypothesis was not supported because there was no evidence that impaired AN threshold distributions were compressed and the growth of AN activity summed across BFs near the stimulus frequency was shallower than normal. Overall, these results suggest that loudness recruitment cannot be accounted for based on summed AN rate responses and may depend on neural mechanisms involved in the central representation of intensity.