Biomedical Engineering Application Brief

Speech Analysis

The Problem

Historically, human speech disorders have often been attributed to psychological rather than physical causes. Until recently, technological limitations prevented thorough investigation of such pathologies as stuttering and spasmodic dysphonia. Dr. Rick Roark and colleagues at the New York Medical College are using a PC-based computer system to gain an understanding of the function of the human central nervous system in normal and pathologic speech production in order to sort out the psychological and physiological aspects of speech problems. The team has had to solve many tough technical problems in order to do so.

Huge Data Files, Complex Analysis

It is difficult to choose instrumentation for a laboratory in any health science field because such research is characterized by huge data sets, complex quantitative measures, and invasive physiologic data acquired from human subjects. Large data sets slow down analysis. Complex data extraction and analysis use a great deal of computer power, resulting in skyrocketing research costs. Data are difficult to reacquire if the first set isn't sufficient. Measuring the body's many movements in speech and analyzing their interactions is difficult, but it is the best way to get an accurate assessment of a problem. By using ingenious measurement techniques on human volunteers and analyzing the resulting gigabytes of data in custom PC computer pro-grams, Dr. Roark and colleagues are creating a picture of a long-misunderstood ailment.

Flexible, User Friendly Software Required

Their research requires an acquisition, data management, display, and analysis system that can perform complex, multi-dimensional data extraction tasks even when data are acquired by diverse instruments at several different sampling rates. The team's software must be flexible, user-friendly, and backed up by superior technical support.