Using computer systems as microswitches for vocal utterances of persons with multiple disabilities.
A simple speech-recognition script can turn any vocal sound into a powerful microswitch that doubles communication attempts.
01Research in Context
What this study did
Two adults with severe motor and developmental disabilities could not speak clearly.
The team built a laptop microswitch that hears any short sound the user makes.
When the software hears the target sound it plays a favorite video or song right away.
Sessions ran during normal day-program activities.
What they found
Both people quickly doubled the number of sounds they made.
The new sounds kept happening after the study ended.
Staff said the clients seemed happier and more alert.
How this fits with other research
Debnath et al. (2023) took the idea further. They added a camera and reached 96 % word accuracy for classroom work.
Northup et al. (1991) did an earlier test. They showed adults with cerebral palsy how to start chats with a communication device. The 2004 study flips the order: sound first, then reward.
Finney et al. (1995) proved that computer speech helps people learn symbols. Connell et al. (2004) show the same boost works when the person is the one making the sound.
Why it matters
If a client can make any noise you can turn that noise into a switch.
Free tools like Windows Speech Recognition let you pick the exact sound and reward.
Try it in five minutes: record the sound, link it to a YouTube clip, and watch the data climb.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Record one client’s grunt in Windows Speech Recognition, link it to a 10-second music clip, and count vocalizations for 15 minutes.
02At a glance
03Original abstract
We assessed the effects of two computer systems used as microswitches for the vocal utterances of an adolescent and a young adult with multiple disabilities. The systems were to respond to three one-syllable utterances of the first participant and nine word-like utterances of the second participant by presenting favorite stimuli linked to those utterances. The computer system used for the first participant relied on a software program developed specifically for this study. The computer system used for the second participant combined a new software program with a commercially available speech recognition program. Results showed that both participants increased the frequencies of their target utterances and obtained high levels of stimulation. Positive performance was maintained at post-intervention checks. Technical and practical implications of the data are discussed.
Research in developmental disabilities, 2004 · doi:10.1016/j.ridd.2003.06.002