Kevin Reilly - US grants

Applicants to the NSF 1991 Minority Graduate Fellowship competition who were awarded "Honorable Mention" status and who enrolled in a computer science or computer engineering graduate program at a U.S. university were eligible to apply to the CISE Directorate for this special award. The purpose of the award is to assist the student in both research and educational activities related to his/her graduate education. The award is made on behalf of the student to the institution with the student's advisor designated as principal investigator.

AbstractContemporary models of the physiological development of speech are based primarily on acoustic observations and orofacial kinematics derived from the vocalizations of infants and toddlers. There is a paucity of corresponding data on the development of vocal control in infants, even though infants exhibit a remarkable range of frequency and intensity modulation. Controlled modulation of these speech parameters entails coordinated activity of the respiratory and laryngeal system, which are quite different in infants and adults. he proposed research project will investigate the development of respiratory and laryngeal control of vocal fundamental frequency (f0) and intensity in groups of six-month olds, one-year olds, two-year olds, and adults. Using procedures that are well-established for adult subjects, changes in f0 and intensity will be elucidated using auditory feedback with perturbed f0, intensity during a vowel imitation task. Intensity changes will also be elicited by changes in the intensity of pink noise presented to the subject. The amplitude and rate of change of f0 and intensity responses will be evaluated for directional sensitivity (i.e., rising versus falling contours) and to reveal when these low-level responses emerge developmentally. Although changes will be elicited in only one parameter at a time (e.g., fo), corresponding changes in the non-target parameter (e.g., intensity) will reveal the extent to which f0 and intensity can e differentially adjusted. In addition to revealing the developmental course for specific autogenic feedback in speech motor control, the proposed investigation will reveal differences in chest wall motion during changes in f0 and intensity. Specific differences in chest wall motion for f0 and intensity modulation will also be evaluated.

DESCRIPTION (provided by applicant): The proposed investigations will elucidate the neural mechanisms underlying the acquisition and representation of speaking skills and other motor learning tasks. Functional magnetic resonance imaging (fMRI) will be used to identify experience-dependent changes in functional activation of brain regions involved in learning an association between the movements of the speech articulators and changes in the frequency/location of a tracking stimulus. An examination of functional activation patterns following subjects' mastery of the task will provide insights into the representation of learned sensory-motor mappings for speech articulatory control. These findings will be used to address several issues regarding the learning and control of goat-directed articulator movements. The identification of statistical correlations between performance measures of learning (e.g., decreases in motor error and experience-dependent patterns of functional activation) will be used to identify brain regions associated with detecting and correction of articulatory motor errors, as well as regions involved in formation and storage of an internal model of the sensory-motor mappings for control of the speech articulators. A between-group comparison of functional activations during auditory vs. visual tracking stimuli will address whether and to what extent the neural processes underlying articulatory motor planning are influenced by the sensory modality of the target space. The random introduction of unanticipated delays in the responsiveness of the tracking stimulus will be used to identify neural structures responsible for processing sensory feedback and detecting discrepancies between predicted and actual feedback. Magnetoencephalography (MEG) will supplement fMRI findings by identifying the temporal sequence of cortical responses to changes in the position/frequency of the tracking stimulus. The results of these analyses will be used to test specific predictions of the DIVA model of speech acquisition and production (Guenther et al., 1995; 1998; in press). Experimental results will be compared to computer simulations of model's learning processes, and any discrepancies between the model and data will be used to guide revisions of the model.

DESCRIPTION (provided by applicant): Speech production is constantly modulated by sensory information from the auditory periphery concerning the accuracy of speech output and ambient noise characteristics. The sensitivity of speech motor output to auditory sensory information has been demonstrated most clearly in studies involving auditory feedback perturbations and changes in background noise signals. However, speech responses observed in these protocols are perturbation are far from uniform and instead depend on a number of characteristics specific to the speech task and speech utterances being investigated. The proposed research examines task and utterance effects on speech responses to auditory feedback perturbations and different background noise signals to evaluate the extent to which the integration of sensory information from the auditory periphery into speech motor commands operates to preserve meaningfully contrastive information in the speech acoustic signal. Investigations of speech responses to perturbations of auditory feedback will examine whether these responses are modulated by the effects of the perturbation on the contrast distance separating the perturbed sound from neighboring sounds. Investigations of speech production in noise will examine the sensitivity of speech responses to different background noise signals during speech task and utterance conditions requiring production of more or less discriminatory information in the speech acoustic signal. The results of this research will provide insight into the question of whether the control of spectral features in speech responses to auditory feedback and background noise share a common control parameter for modulating discriminatory power in speech output. PUBLIC HEALTH RELEVANCE: The purpose of this study is to investigate how the motor commands for speech production are integrated with sensory information from the auditory periphery to preserve meaningfully contrastive information in the speech acoustic signal in different speaking conditions. The proposed work will supplement the existing literature on sensorimotor control for speech and provide a more comprehensive account of the role of auditory sensory information in speech communication.