Complex-tone pitch representations in the human auditory system

Understanding how the human auditory system processes the physical properties of an acoustical stimulus to give rise to a pitch percept is a fascinating aspect of hearing research. Since most natural sounds are harmonic complex tones, this work focused on the nature of pitch-relevant cues that are necessary for the auditory system to retrieve the pitch of complex sounds.

The existence of different pitch-coding mechanisms for low-numbered (spectrally resolved) and high-numbered (unresolved) harmonics was investigated by comparing pitch-discrimination performance across different cohorts of listeners, specifically those showing enhanced pitch cues (i.e., musicians) and those typically having disrupted pitch cues (i.e., hearing-impaired listeners).

In particular, two main topics were addressed: the relative importance of resolved and unresolved harmonics for normal-hearing (NH) and hearing-impaired (HI) listeners and the effect of musical training for pitch discrimination of complex tones with resolved and unresolved harmonics. Concerning the first topic, behavioral and modeling results in listeners with sensorineural hearing loss (SNHL) indicated that temporal envelope cues of complex tones with unresolved harmonics may be enhanced relative to NH listeners at the output of peripheral auditory filters.

This enhancement of temporal envelope coding was found to be ascribed to a reduction of cochlear compression. Since frequency selectivity and temporal fine structure (TFS) cues are known to be degraded in listeners with SNHL, it is likely that HI listeners rely on the enhanced envelope cues to retrieve the pitch of unresolved harmonics.

Hence, the relative importance of pitch cues may be altered in HI listeners, whereby envelope cues may be used instead of TFS cues to obtain a similar performance in pitch discrimination to that of NH listeners. In the second part of this work, behavioral and objective measures of pitch discrimination were carried out in musicians and non-musicians.

Musicians showed an increased pitch-discrimination performance relative to non-musicians for both resolved and unresolved harmonics, although their benefit was larger for the resolved harmonics. Additionally, task-evoked pupil responses were recorded as an indicator of processing effort while listeners performed a pitch-discrimination task.

Although the difficulty of the task was adjusted for each participant to compensate for the individual pitch-discrimination abilities, the musically trained listeners still allocated lower processing effort than did the non-musicians to perform the task at the same performance level. This finding suggests an enhanced pitch representation along the auditory system in musicians, possibly as a result of training, which seemed to be specific to the stimuli containing resolved harmonics.

Finally, a functional magnetic resonance imaging paradigm was used to examine the response of the auditory cortex to resolved and unresolved harmonics in musicians and non-musicians. The neural responses in musicians were enhanced relative to the non-musicians for both resolved and unresolved harmonics in the right auditory cortex, right frontal regions and inferior colliculus.

However, the increase in neural activation in the right auditory cortex of musicians was predictive of the increased pitch-discrimination performance only for resolved harmonics. These results suggest a training-dependent effect in musicians that is partially specific to the resolved harmonics.