The neural basis of speech perception

Abstract

The standard view of the functional neuroanatomy of speech processing attributes the neural basis of speech perception to the left posterior superior temporal cortex, in line with the doctrine that a number of areas in the left hemisphere are specialized for the representation and processing of speech and language. The research presented in this dissertation reevaluates this model. By systematically distinguishing between the neural basis for speech perception proper (construed as one interface with the language processing system and hence a subroutine of auditory language comprehension) and the neural basis for other aspects of language representation and processing, it is demonstrated that both the left and right superior temporal gyri (non-primary auditory cortex) are implicated in the construction of speech sounds. Evidence that speaks to the question of the neural basis of speech perception is drawn from three sources: the deficit-lesion correlation literature on an auditory verbal agnosia called pure word deafness (Chapter 2), the literature on brain activity measurements using metabolic functional imaging methods such as positron emission tomography (Chapter 3), and experimental results derived from the new electromagnetic brain imaging technique magnetoencephalography (Chapter 4). Motivated by the observations from the deficit-lesion and the neuroimaging literature, the thesis develops the model that speech perception is mediated by the posterior superior temporal gyri in both hemispheres. Two hypotheses about the speech perception system are experimentally investigated. The first hypothesis is that one can reliably detect left and right supratemporal cortical activation generated by speech sounds, and the second that the execution of task demands differentially affects the two hemispheres. These two questions were evaluated in three magnetoencephalography studies recording the auditory evoked neuromagnetic field generated by tones, vowels, and syllables. In passive listening conditions, the evoked response data as well as the source localization data revealed no systematic differences between left and right auditory association cortices, supporting the model that both areas contribute to the processing of speech sounds. The comparison between a task-driven and a passive condition showed that the right and left auditory association cortices responded differently to attentional requirements: compared to a passive listening condition, the attended condition yielded simultaneous response amplitude increases in the left and decreases in the right in the same supratemporal cortical area. This research concludes that both left and right posterior superior temporal gyri play significant roles in the processing of speech sounds, although the precise nature of the computations remains unspecified. (Copies available exclusively from MIT Libraries, Rm. 14-0551, Cambridge, MA 02139-4307. Ph. 617-253-5668; Fax [phone].)