haskins logo
Applications of research on speech production

A long-term effort at Haskins Laboratories has been to study speech production, a fundamentally human process and the foundation of interpersonal communication. Although human speech is essential for individuals to function and thrive in today’s society, it is estimated that one out of every six people suffers from some form of communication disorder during the course of life (National Institute for Deafness and Other Communication Disorders, National Institutes of Health). The social, emotional, educational, economic and vocational consequences for individuals and society are significant. An even heavier and more significant loss can be attributed to the diminished productivity of those who suffer communicative disorders. Although speech is ubiquitous in our everyday lives, and it is one of our most important human functions, we still have at best an incomplete understanding of how the process works and how or why it malfunctions in many speech and language disorders.

At Haskins Laboratories our approach to speech production has been to focus on developing a detailed understanding of the normal processes of speech, from the movements producing the sounds to the brain processes that control them. At the level of the individual speech movements, producing the sounds of the language involves as many as 100 muscles, necessitating smooth and accurate coordination to produce intelligible speech. Current research combining direct recordings of movements of the speech articulators with articulatory speech synthesis has provided theoretical and practical insight into the complex coordination of movements necessary to produce speech signals. Nonlinear dynamical models of self-organizing systems have been applied to elucidate principles of coordination and regulation among speech gestures. These task-dynamic models are focused on understanding the manner in which speech gestures both cohere in higher order units during normal speaking and reorganize spontaneously in phenomena such as speech errors. Most recently, this dynamical framework has served as the basis for modeling the temporal aspects of speech’s supra-gestural prosodic or expressive structure.

One important extension of our research has focused on studying how young children who are normally learning to speak acquire the skills to control and coordinate vocal tract gestures. Through research on babbling in infants from a variety of language environments, we are learning how infants coordinate structures of the vocal tract and how and when that coordination begins to reflect the special characteristics of the language to which the child is exposed.

As for speech disorders, they can be better understood when normal speech processes are known. Stuttering, for example, a developmental disorder in the fluent production of speech, is a motor disorder with significant neurophysiological manifestations. This requires a research approach that focuses not only on the behavior (the symptoms) but also on the mechanisms underlying them (the cause). We are integrating our theoretical perspective on normal speech production processes with functional neuroimaging studies of normal brain processes to develop a neurobiological model of speech. This approach allows us to address hypotheses about the brain regions and neural interactions that significantly contribute to the breakdown in speech fluency that is characteristic of stuttering.

In contrast to stuttering, where the neural mechanisms are still unknown, dysarthria, a collective name for a group of speech disorders, damage to neural systems that lead to disturbances in muscular control over the speech articulators. Dysarthria includes problems in oral communication due to paralysis, weakness, or incoordination of the speech musculature. Here we are applying our technical and technological sophistication to describe and detail the speech movement and perceptual consequences of brain damage in patients with Parkinson’s Disease. Using instruments, such as Ultrasound and Electromagnetic Articulography. we are able to observe the movements of the speech articulators (the tongue, jaw, and lips) directly while patients produce speech. By combining these observations with the perceptual evaluation of patients’ speech, we are able to understand the specific speech motor problems associated with their disease. As with all our research applications, the information fundamental to diagnose and efficiently treat a wide range of speech and language disorders derives from investigations of the normal speech production process.