Speech-Language Pathology/Stuttering/Computers and Electronics for Fluency-Shaping

The associative stage of motor learning requires feedback. In sports this is called knowledge of results. E.g., in golf or tennis you see where the ball goes after you hit it. Playing golf or tennis on a dark, foggy night would be impossible.

Feedback quality is affected by speed. If you hit ten golf balls on a dark, foggy night, then the next day find one of the balls 150 yards away, you'll have no memory of what you did right to hit it so far.

Feedback quality is also affected by accuracy. If you and your buddy each hit a golf ball, and one ball goes 150 yards but you don't know whose ball it was, you have inaccurate feedback.

Or the observer gets bored. If you hit golf balls for hours, and have a person telling you how far the balls go, sooner or later the person will stop paying attention.

Which Fluency Skills Need Feedback?
When you're learning fluent speech motor skills, you need knowledge of results. Some skills are easy to observe. E.g., resting your hand on your stomach tells you whether you're using diaphragmatic (relaxed) breathing or thoracic (speech) breathing.

Your articulators (lips, jaw, and tongue) are a little harder to be aware of, as you can't see them. But you have good proprioceptive awareness of these muscles, so developing awareness and control isn't hard.

Your vocal folds are another story. These muscles are deep in your throat. You can't touch them or see them. Most people don't even know they have vocal folds.

The most difficult feedback is with the timing of all this. E.g., your speech-language pathologist tells you to exhale a little air and then increase your vocal fold tension. You do this slowly and correctly. Then she tells you to increase the speed. You must execute these movements within hundredths of a second. You can't tell whether you're doing it right, and most speech-language pathologists can't either.

A fluency specialist who's helped hundreds of stutterers has better-trained ear and visual skills and gives better quality of feedback than a speech-language pathologist who's never treated a stutterer.

Biofeedback Devices
Biofeedback is the measurement and display (to the user) of a physiological activity, to enable the user to improve awareness and control of the activity.

Biofeedback machines provide:

 Faster, more precise, and more reliable feedback than a human observer. Machines can provide feedback in real-time, beeping the instant you make a mistake. Machines can accurately measure things humans can't see or hear. And machines never get bored, even after hours of practice. With the computer training the physical speech motor skills, the speech pathologist can spend more time on the psychological aspects of stuttering. If you learn visually rather than aurally, you may learn faster with the computer displays than by listening to your speech on a tape recorder. Some devices are designed for home practice use as well as clinical use. 

But biofeedback devices aren't a &quot;miracle cure&quot; for stuttering. Stuttering therapy works on many motor skills, so you might need several devices. And you still need a speech-language pathologist to train you to do the target motor skills (cognitive stage). The machines can only help you to refine your skills (associative stage).


 * Wikipedia article about biofeedback

CAFET and Dr. Fluency
The Computer-Aided Fluency Establishment and Trainer (CAFET) and Dr. Fluency are computer-based stuttering biofeedback systems. Both use a microphone to monitor vocal volume, as a surrogate for vocal-fold activity, and a chest-strap to monitor breathing. Dr. Fluency uses two chest straps, to differentiate diaphragmatic (relaxed) breathing from thoracic (upper chest) breathing.

You see your breathing and vocal volume displayed on the computer screen, along with instructions or error messages. The two computer systems train similar speech motor skills:

<ol> <li>Relaxed, diaphragmatic breathing.</li> <li>Continuous breathing. The computer alerts you if you hold your breath more than 1/3 of a second.</li> <li>Gradual exhalation, as opposed to the rapid, uncontrolled exhalation associated with stuttering.</li> <li>Pre-voice exhalation, or letting a little air out before you begin tensing your vocal folds.</li> <li>Gentle onset, or gradually increase vocal volume. The computer alerts you if your vocal volume changes too rapidly. The computer also alerts you if your voice is too quiet for your air flow (which sounds breathy).</li> <li>Continuous phonation. Breaks in vocal volume are shown on the computer monitor.</li> <li>Adequate breath support. The computer alerts you if you continue to talk after the point at which you should take another breath.</li> <li>Phrasing. Each of the above seven speech targets is taught first with vowels, then progressing to monosyllabic words, then to marked-length phrases.</li> </ol>

An unpublished study of the CAFET program with 197 adults and teenagers reported that 82% met fluency criteria six months after completing the program; 89% were fluent after twelve months; and 92% were fluent two years post-therapy.

Electromyography (EMG)
EMG measures muscle activity via electrodes taped to your skin. Several studies found EMG to be effective for treating stuttering. These include the large study that found computers to be more effective than speech-language pathologists.

EMG surface electrodes tape to your skin. They are excellent for monitoring your masseter (jaw muscle) or orbicular oris (lips); pretty good for monitoring your tongue or respiration muscles; and useless for monitoring your vocal folds. Your vocal folds are deep in your neck, surrounded by other muscles. A single-channel EMG is usually set up with one electrode on the jaw and the other electrode at the base of the throat, which monitors all muscle activity from thoracic respiration to articulation. This set-up monitors all speech-production muscles (including vocal fold activity) at once. When the red light goes on (indicating high muscle activity) it may not be apparent which muscle is tense. Also smiling and swallowing are indicated as high muscle activity. These problems could be avoided by using a multichannel EMG, but those are costly and time-consuming to set up.

Vocal Frequency Biofeedback
The CAFET and Dr. Fluency monitor vocal fold activity indirectly via vocal volume, which assumes that a quiet voice means relaxed vocal folds, and loud voice means tense vocal folds. Another method is to monitor the frequencies of the stutterer's voice. Relaxed vocal folds generally produce a deep-pitched voice; tense vocal folds generally produce a high-pitched voice. Vocal volume and vocal frequency biofeedback use a microphone; no electrodes are needed.

Kay Elemetrics' Visi-Pitch device monitors both vocal volume and vocal frequency, separately. Casa Futura Technologies has devices that monitor a combination of vocal volume and frequency.