Laurel H. Carney
This presentation will focus on information contained in the temporal patterns of auditory nerve (AN) responses and its quantification in terms of its ability to explain performance on a sound-level discrimination task. The temporal patterns of responses of AN fibers include phase-locked responses to simple and complex stimuli. The phase of phase-locked responses varies with sound level due to nonlinear properties of mechanical tuning in the inner ear. Colburn (1982) used optimal decision theory to study the potential for average discharge rate and synchrony (strength of phase-locking) to explain sensitivity (d') in a level discrmination task. We have extended this approach to include nonlinear phase cues, using a simple analytical model for AN responses that includes phase shifts with sound level. Our results show that there is significant information contained in the relative timing of discharges across AN fibers that are tuned to different frequencies. Unlike information contained in the average rates and synchrony of single fibers, or even populations of fibers, the information derived from nonlinear phase cues persists over a wide range of levels. The derivation for the sensitivity index (d') that includes the nonlinear phase cues shows that this cue is weighted (in an intuitively reasonable way) by both the discharge rate and synchrony. The combination of these three response properties thus provides information relevent for discrimination in a form that can be processed by a single physiologically realistic mechanism, coincidence detection.
In addition, I will present a more recently described property of AN responses: frequency glides in the impulse responses of AN fibers change in size and direction as a function of characteristic frequency. The frequency glide is an apparently linear property of peripheral tuning; however, it interacts with the nonlinear tuning to create frequency shifts in tuning as a function of sound level. This property will be introduced and discussed in terms of its implications for temporal coding of complex sounds.