Pitch of Harmonic Complex Tones: Rate Coding of Envelope Repetition Rate in the Auditory Midbrain.

Yaqing Su,Bertrand Delgutte

doi:10.3813/aaa.919239

Abstract

SummaryEnvelope repetition rate (ERR) is an important cue for the pitch of harmonic complex tones (HCT), especially when the tone consists entirely of unresolved harmonics. Neural synchronization to the stimulus envelope provides a prominent cue for ERR in the auditory periphery, but this temporal code becomes degraded and gives way to rate codes in higher centers. The inferior colliculus (IC) likely plays a key role in this temporal-to-rate code transformation. Here we recorded single IC neuron responses to HCT at varying fundamental frequencies (F0). ERR was manipulated by applying different inter-harmonic phase relationships. We identified a subset of neurons that showed a ‘non-tonotopic’ rate tuning to ERR between 160 and 1500 Hz. A comparison of neural responses to HCT and sinusoidally amplitude modulated (SAM) noise suggests that this tuning is dependent on the shape of stimulus envelope. A phenomenological model is able to reproduce the non-tonotopic tuning to ERR, and suggests it arises in the IC via synaptic inhibition.

Highlights

Harmonic complex tones present in speech, musical sounds and animal vocalizations evoke a strong pitch sensation at their fundamental frequency (F0), even if they contain no energy at F0 (“missing fundamental”)
Studies of the auditory nerve (AN) and cochlear nucleus (CN) have described multiple potential codes to pitch cues, including a rate-place code for resolved harmonics, temporal codes based on interspike interval distributions, and spatio-temporal codes dependent on both cochlear frequency selectivity and neural phase locking [1, 2, 3, 4]
The rate profile shows multiple peaks occurring when the F0 is a small integer submultiple of the best frequency (BF) (BF, BF/2, BF/3, ...), reflecting cochlear tuning to resolved harmonics

Summary

Introduction

Harmonic complex tones present in speech, musical sounds and animal vocalizations evoke a strong pitch sensation at their fundamental frequency (F0), even if they contain no energy at F0 (“missing fundamental”). Pitch plays important roles in speech and music perception and in the perceptual organization of auditory scenes, the neural mechanisms underlying pitch perception are still poorly understood. Studies of the auditory nerve (AN) and cochlear nucleus (CN) have described multiple potential codes to pitch cues, including a rate-place code for resolved harmonics, temporal codes based on interspike interval distributions, and spatio-temporal codes dependent on both cochlear frequency selectivity and neural phase locking [1, 2, 3, 4]. IC, the major nucleus in the auditory midbrain, is a logical target for addressing this question because it is the site of a transformation from a temporal code to a rate code for the frequency of amplitude modulations [6, 7]

Methods

Results

Discussion

Conclusion