SPEECH SOUND PROCESSING IN AUDITORY CORTEX

DESCRIPTION: (Adapted from the Applicant's Abstract.) The overall objective of this project is to employ electrophysiological responses to human speech sounds within monkey primate auditory cortex (AI) as a model of neural mechanisms that underlie cortical processing of acoustic and phonetic features of speech. Using multicontact electrodes, three complementary techniques will be used to examine the activity from neuronal ensembles; multiple unit activity (MUA), auditory evoked potentials (AEPs) and the derived current source density (CSD). CSD analysis delineates the temporal and laminar distribution of the current sources and sinks that reflect net synaptic excitation and inhibition whereas phasic MUA patterns provide information on changes in the net firing rate of neurons in the vicinity of the recording electrodes. The neuronal ensemble responses will also be compared to single unit recordings to evaluate relationships between these procedures. These studies will define the obligatory auditory cortical evoked responses elicited by speech stimuli and form the basis for future comparisons with responses associated with discriminative task requirements, Speech sound processing will be defined in terms of the spatio-temporal pattern theory of complex sound encoding, which states that complex sounds are uniquely encoded in the overall response patterns generated within auditory cortex. Thus, the different response patterns of MUA and transmembrane currents elicited by synthetic consonant-vowel (CV) syllables that differ in their specific vowels, consonant place of articulation and onset of voicing will be defined and related to the acoustic parameters of the stimuli and to the laminar, tonotopic, binaural column and ampliotopic organizations of AI. Response differences between thalamocortical fibers and cortical cells within different laminae will be compared to delineate the transformations that occur in the response at successive levels of cortical processing. Modulation of the response patterns induced by formant interactions and by changes in stimulus intensity will be assessed. The applicability of findings for synthetic speech sounds to naturally produced syllables will be determined. The relative importance of changing formant transitions versus constant formant onsets in the neural encoding of consonants will be assessed. The hypothesis that the discrete encoding of the onsets of the unvoiced and voiced portions of stop CV syllables defines the psychoacoustic boundary demarcating the perception of these two classes of consonants will be tested. Finally, the relationship of the intracortical responses to AEPs recorded over the cortical convexity will be determined to define those aspects of the intracortical processing that are reflected at the cortical surface. These relationships will serve as a model for the neural events that underlie the scalp recorded human AEPs to speech, and in so doing, will increase our understanding of both normal and pathological speech processing.