This proposal seeks to better understand the central nervous system control (CNS) of the larynx in voice production. Because certain important analysis techniques cannot be applied to humans, the study of model systems is required. Study of two such model systems is proposed. First, single-units will be recorded in CNS structures in cats instrumentally conditioned to vocalize. Recordings are proposed for two CNS targets: 1) regions in the basal ganglia; and 2) ventrolateral thalamus. These are structures at mid-levels of the proposed CNS control system. They have been implicated in human voice production by clinical studies of brain lesions, and in animal voice control by lesion and electrical stimulation studies, but data concerning how they are involved, are lacking. A second model proposed for study is a computational model of CNS laryngeal control. The model is a hybrid consisting of two parts. The most peripheral portion is a mathematical model of laryngeal function, based on biomechanical principles, whose study is of interest in its own right. The representations of muscles in this mathematical larynx are, in turn, controlled by a trainable artificial neural network model of the CNS laryngeal control pathways. A goal of the proposal is to constrain the network model as much as possible by known anatomy, physiology, and vocal behavior so that, after training, its function comes to match the real laryngeal control system. Once it has been developed, the model can be subjected to various manipulations, such as the equivalent of CNS lesions, and the effects on voice production by the model studied. The combined study of these two model systems is relevant to several health issues in communication disorders. The data from the mammalian system will clarify the functional role of some mid-level CNS structures in normal voice control. In addition, the data are needed to aid the process of model development. Successful development of the computational model, in turn, provides a powerful means for studying normal and abnormal mechanisms underlying voice production. It also may prove important in control of laryngeal prostheses, and in reducing animal experimentation.
| Farley, G R (1996) A biomechanical laryngeal model of voice F0 and glottal width control. J Acoust Soc Am 100:3794-812 |
| Farley, G R (1994) Control of voice F0 by an artificial neural network. J Acoust Soc Am 96:1374-9 |
| Farley, G R (1994) A quantitative model of voice F0 control. J Acoust Soc Am 95:1017-29 |
