A recent study got me thinking about how best to support my non-verbal students. The study deals with non-fluent aphasia. Non-fluent aphasia is a language disorder that makes it difficult for individuals to speak in complex sentences. Symptoms include using fewer verbs, pronouns, and prepositions. It's been presumed these result from a brain deficit in language processing.
However, the new study refutes this theory. Researchers analyzed spoken and written samples from people with and without non-fluent aphasia. Using AI, they found patients compensate for simple sentences by choosing more informative, semantically richer words to sustain sentence information.
Rather than being caused by brain deficits, the symptoms appear to be an adaptive response to difficulties with complex sentences. This challenges the common understanding of non-fluent aphasia's origins.
The findings indicate revisions may be needed in treatment approaches. Current methods target presumed defective symptoms, but patients are already adaptively using informative words. Treatment could be optimized by promoting this compensation strategy further. AI could help identify the most informative words to suggest given context.
The compensation strategy of using AI to suggest informative words could potentially benefit some non-speaking autistic individuals as well, for a few reasons:
Like those with non-fluent aphasia, non-speaking autistics face challenges with producing complex sentences and spoken language. AI word suggestions could aid communication.
Non-speaking autistics often use augmentative and alternative communication (AAC) devices. AI could be incorporated into AAC technology to offer word recommendations that maintain sentence context and meaning.
Autistics tend to have enhanced abilities to process semantic and contextual information. Leveraging these strengths through AI-powered word suggestions could improve communication flow.
Some non-speaking autistics have motor planning challenges analogous to apraxia of speech. AI aids could support their visual language and writing.
However, the benefit would depend on the individual's abilities and difficulties. AI, of course, should not replace the communicator's autonomy. And comprehensive augmentative communication evaluation is still essential. But with proper design, AI could potentially enhance communication aids for select non-speaking autistics in a supportive role. Of course, more research is needed on appropriate implementation.
Fascinating!