Cortical codes for pitch and its role in selective listening

Thesis

Pitch is a fundamental feature of sound that underlies our perception of speech and music, and enables us to segregate sound sources in complex acoustic environments. Despite this essential role, it is unclear how pitch is represented in the brain. Single neurons that selectively encode pitch have been identified in marmosets, but studies in other species, including humans, have instead identified distributed pitch codes that are often specific to acoustic structure. In this thesis, I addressed these conflicting findings through electrophysiology recordings in ferret auditory cortex. I found that a subset of neurons represented pitch based on harmonic spacing, while another subset derived pitch from temporal periodicity. A third subset invariantly represented pitch across both cue classes, the first evidence of such neurons outside of marmosets. Each cue class evoked a different population code for pitch, and population responses at sound onset and offset were equally and independently informative about pitch. I found that cortical representations of the pitch of complex sounds did not systematically relate to pure tone frequency encoding, either in single neurons or population codes. In a novel two-tone streaming paradigm, I showed that ferrets can use the pitch of both pure tones and complex sounds to segregate competing auditory streams, the first demonstration of pitch as a cue for selective listening outside of humans. Pitch perception and selective listening often worsen with ageing, even when audiograms reveal only mild hearing impairment, and improved clinical interventions will depend on a better understanding of the neural mechanisms underlying these fundamental aspects of hearing. The cortical codes for pitch I have described in ferrets unites diverging findings across several species, and the novel auditory streaming paradigm I designed opens the door for future studies to use invasive techniques only possible in animals to identify the neural mechanisms underlying selective listening based on a variety of acoustic cues.

Authors: Tarka, V

• DOI: 10.5287/ora-5za65oare
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: ferret; pitch; electrophysiology; behavioural neuroscience; auditory cortex; computational neuroscience
• Subjects: auditory neuroscience; neural physiology