Extending the evaluation of a computer system used as a microswitch for word utterances of persons with multiple disabilities.

The team built a computer microswitch that listens for target words. When it hears one, it plays a short clip of the speaker’s favorite music or video.

Three adults with multiple disabilities used the system during daily 15-minute sessions. Staff recorded how often each person said the target word.

The study ran for several weeks, then checked again two months later to see if the gains stuck.

All three participants said the target word more often once the computer started giving them preferred clips.

The increase stayed in place at the two-month follow-up, showing the effect was not just a quick blip.

Recognition was not perfect; the computer sometimes missed words or gave clips for the wrong word, yet talking still went up.

Lancioni et al. (2011) used the same microswitch-plus-stimuli idea, but for choosing items instead of talking. Both studies show the tech works for different responses, so you can swap the target behavior without rebuilding the whole setup.

Kennedy et al. (1993) first proved that a simple switch giving kids their favorite toys boosted alertness during peer play. The 2005 paper upgrades that switch into voice-recognition software, moving the field from pressing to speaking.

Austin et al. (2015) later added posture control: preferred clips paused if the client slumped. Their twist shows the computer can handle two jobs at once—talking and sitting up—so you do not have to pick just one goal.

If you have clients who rarely speak, a voice-detecting microswitch can turn any word into automatic reinforcement. Start with one highly preferred video and one easy word. Track hits and misses for a week, then fine-tune the recognition settings. Even imperfect detection can still grow language.