Brainstem Coding of Pitch – the Butte Hypothesis

Philip X. Joris

Pitch has been described as “the heart of hearing theory” (Plomp 2002), but its physiological underpinnings are unclear and have been much debated in preceding ISH meetings. I propose a new view in which a brainstem property plays a key role in the coding of periodicity. I refer to this property as “entracking”, a contraction of entrained phase-locking. A robust code for pitch exists in the auditory nerve in the form of a pooled interspike interval (ISI) distribution (Cariani and Delgutte 1996), which resembles the stimulus autocorrelation function. An unsolved question is how such a representation could be grounded in physiology, as its computation seems to require a combination of coincidence detectors and a range of delays extending over many milliseconds. I propose a scalar rather than vector code of periodicity by virtue of coincidence detectors that code the dominant ISI directly into spike rate through entracking. Perfect entracking means that a neuron fires one spike per repetition period, so that the firing rate equals the repetition frequency. Key properties of entracking are that it improves with increasing SPL and that the number of entracking neurons increases with SPL. The main limitation in this code is the upper limit of firing (usually near 500 Hz). It is proposed that entracking provides a periodicity tag which is superimposed on a tonotopic analysis. At low SPLs and fundamental frequencies > 500 Hz, a spectral or place mechanism codes for pitch. With increasing SPL the place code degrades but entracking improves and first occurs in the neurons with the lowest thresholds for the spectral components present. The prediction is that populations of neurons extended across CF form plateaus (“buttes”) of firing equaling the fundamental frequency. The core idea is that entracking brings a stratification to firing rate tied to periodicity.