The proposed research addresses hypotheses concerning the way speech movements are planned and executed. The hypotheses are guided by a computational model of the neural control of speech movements combined with a biomechanically based vocal-tract model. To test these hypotheses, we conduct experiments with speakers and listeners in which we measure articulatory movements, speech acoustics, perception, and brain activation. We manipulate speech condition (e.g., clear vs. fast speech), phonemic context, and speech sound class, and we introduce transient and sustained perturbations. We also perform simulation experiments in which we adapt the vocal-tract model to the morphologies of individual speakers. Movements of this vocal-tract model are commanded by our neural control model for the same utterances and conditions used in our experiments, and the resulting movements and acoustics are compared to those of the modeled speaker, in order to test the model. The research addresses 3 inter-related issues: 1. The nature of phonemic and syllabic goals for articulatory movements. Here we study the effects of speaking conditions and speaker differences on auditory goals for vowels; the effects of speakers' discriminative capacities on the phoneme contrasts they produce; and the effects of auditory and articulatory perturbations on phonemic contrast and their interaction with sound class and speaker acuity. 2. The role of feedback and feedforward mechanisms in the control of speech movements. The performance of these control sub-systems is probed by introducing unpredictable and sustained perturbations. We test model-based hypotheses about changes in brain activation and acoustic and articulatory trajectories elicited by each of these interventions. 3. Movement trajectory planning in the concatenation of phonemes and larger units. In this study we investigate the planning of movements in the concatenation of words. We measure articulatory movements in gestural overlap and their acoustic and perceptual correlates in order to test the hypothesis that such overlap leaves perceptible acoustic cues for listeners. We believe these studies will provide important information regarding the neural control mechanisms underlying speech production. ? ?
Showing the most recent 10 out of 19 publications
