The overall objective of this project is to elucidate neural mechanisms underlying cortical processing of speech by examining electrophysiological responses in monkey primary auditory cortex (A1). Structural anomalies in auditory cortex, and perceptual abnormalities in the processing of speech and other sounds, occur in people with developmental dysphasia. Relating these deficits to dysfunction of specific neural events requires an understanding of normal cortical processes that is best afforded by intracranial recordings. Many features of human phonetic perception occur in monkeys, indicating that these animals are a reasonable animal model for studying cortical responses to speech. Using multicontact electrodes, three complementary measures will examine the activity from neuronal ensembles, multiunit activity (MUA), evoked potentials (AEP) and the derived current source density (CSD). CSD delineates temporal and laminar distributions of current flow that reflect net synaptic excitation and inhibition, whereas MUA patterns define changes in the firing rate of neuronal ensembles. Through their relationship with the AEP, intracortical responses can be directly linked with homologous responses in humans. These procedures yield stable measures of the synchronized neural activity required for complex sound encoding. This project will clarify two key problems in phonetic perception: (1) the relationship between the acoustic signal of a speech sound and its phonetic representation and (2) how speech sounds are categorically perceived. These problems will be addressed by testing four hypotheses relating to the mechanisms underlying encoding of consonant place of articulation and voice onset time (VOT): (1) VOT is categorically encoded by temporal response patterns evoked by acoustic transients occurring at consonant onset and the onset of syllable voicing, (2) mechanisms of VOT encoding can be modeled through the use of simplified tone burst stimuli, (3) place of articulation is categorically encoded by responses to the onset spectra of consonants, and (4) place of articulation encoding in A1 is improved when dynamic spectral features embedded in the formant transitions are utilized. Testing these hypotheses will require that acoustic parameters of the stimuli be related to organizational features of A1, demonstrating that systematic changes in categorical response boundaries occur in parallel with stimulus modifications, and that prolonged formant transitions increase differences by increasing the duration of spectral disparity between the syllables. These studies will define normal mechanisms of A1 speech processing, and serve as a benchmark to evaluate dysfunctional mechanisms associated with abnormal language and language development.
Showing the most recent 10 out of 29 publications
