Autocorrelation model of periodicity coding in bullfrog auditory nerve fibers

Abstract

Behavioral evoked vocal responses of male bullfrogs (Rana catesbeiana) are sensitive to slight disruption of the 1 st harmonic periodicity in synthetic calls produced by changes in component phase. We implemented a computational model that estimates stimulus period from autocorrelation functions (all-order interval histograms) of auditory nerve responses aligned in time across frequency channels. Output of the model shows differences between stimuli in a manner reflecting behavioral data. In contrast, peri-stimulus-time histograms aligned across channels do not clearly reflect behavioral responses. The estimate of stimulus period derived from time-alignment autocorrelation shown here is similar to display of pitch images in models based on mammalian data. © 2000 Acoustical Society of America.