Use this URL to cite or link to this record in EThOS:
Title: Hierarchical organisation of voice and voice gender perception
Author: Charest, Ian
Awarding Body: University of Glasgow
Current Institution: University of Glasgow
Date of Award: 2010
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Restricted access.
Access from Institution:
The most important sound in our auditory environment is the human voice. Voice professionals, whether they are teachers, radio hosts, sport coaches, use their voice on a everyday basis to earn their living and communicate information and knowledge. We grow up spending most of our time everyday listening to voices in school, at the sports club, on t.v., etc. So much that by the time we are adults, voice plays a major role in our everyday social interactions. Yet, while extensive research has been conducted on speech perception voice alone has only just started generating more and more interest in the cognitive neuroscience research community. Voice is not "just" a speech carrier, it conveys rich paralinguistic information such as gender, age, identity or affective state. A theoretical model which emphasises the similarities between face and voice processing was recently introduced, suggesting a serial and parallel processing pathway of voice information leading to high level cognitive processes like person identification. Globally this model of voice processing suggests an extraction of low-level acoustic features, followed by a voice structural encoding leading to parallel pathways for the recognition of speech, affect and identity related information. Furthermore, this model suggested potential interactions with face perception pathways. In this thesis, I investigated two different stages of this voice perception model. First, little is known about the speed at which the distinction between vocal and non-vocal sounds is performed, i.e. is there a time-frame where the "voice structural analysis" would occur. Using electroencephalography, we conducted an experiment in order to delineate this voice vs. non-voice perception time-frame. I observed an early electro-physiological response preferential to voice stimuli, emerging around 164 ms on fronto-temporal electrodes FC5 and FC6 which was termed the "fronto-temporal positivity to voice". Second, little is known about the neural basis behind the perception of paralinguistic information such as identity, gender or affective state contained in the human voice. I used voice gender as a tool to investigated the "voice recognition units" stage of the voice perception model. The cognitive processes behind voice gender perception are still under debate, and more precisely, the nature of the representation of voice gender, whether it is organised around low level acoustical discriminants, or relies on high level categorical representations still remains unclear. Voice gender continua can be created in order to parametrically control the degree of gender contained in voice. I investigated the importance of low level acoustic features using the recently developed auditory morphing algorithms. I averaged 32 male and 32 female voices in order to "approximate" a prototypical voice for each gender. From those prototypes, I generated caricatures by exaggerating the acoustical properties of the male prototype in reference to the female prototype. Those voice composites were included along with 3 pairs of male and female voice exemplars in a voice gender adaptation experiment. I observed significantly stronger perceptual after-effects caused by adaptation from the voice gender caricatures. This result provides evidence for a determinant role of the low level acoustical features in our ability to perceive the gender of a voice. Finally, using functional magnetic resonance imaging (fMRI), I investigated whether brain regions of the auditory cortex are sensitive to voice gender, voice gender adaptation, and whether a dissociation between extraction of acoustical features and higher level, perceptual representations could be achieved. I used voice gender continua and an event-related fMRI design called the continuous carry-over design to assess these working hypotheses. I observed a covariation between BOLD signal and the degree of acoustical differences in consecutive voices in the anterior part of the right superior temporal sulcus, where the extraction of voice gender related acoustical features occurs. Furthermore, I observed a higher level network involving the bilateral inferior frontal gyrus, the insula and the anterior cingulate cortex where a summary of acoustical features would be input from auditory areas enabling a voice gender categorisation.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: BF Psychology