Towards Text-based Augmentative Communication in Individuals with Intellectual Disabilities and Difficult-to-Understand Speech

Abstract

Individuals who have difficult-to-understand speech and minimal reading skills are often limited to icon or picture-based augmentative and alternative communication (AAC) strategies. However, basic phonemic awareness skills could make strategies such as alphabet supplementation, in which the speaker selects the first letter of words they are speaking, a viable AAC option for increasing the extent to which the speaker's words are understandable to listeners (i.e., speech intelligibility). We conducted two studies with adults with severe intellectual disabilities, difficult-to-understand speech, and limited reading skills. The purpose of Study 1 was to teach participants to select the onset letter of a large number of spoken words. Six phonemes were targeted for instruction, and discrimination of onset phonemes was trained by pairing words with the same rime (e.g., mall/call) in a computerized matching-to-sample task. We also assessed if training the onset phoneme in isolation or with multiple word pairs would result in generalization to untrained spoken words beginning with the trained onset phonemes. Participants learned to select the onset letter of more than 60 words. However, consistently high accuracy in tests of generalization was not observed, even as an increasing number of exemplars were trained. In Study 2, our goal was to assess and then train the component skills to use an augmentative keyboard in a contrived communication task. After establishing picture naming responses to evoke speech and onset-letter selection with 30 new words, we assessed the effects of the alphabet supplementation strategy on participants' speech intelligibility. Results showed that listeners understood twice as many words when the augmentative keyboard was used to indicate the first letter of the words being spoken.