About 2 million Americans live with aphasia. The challenge is harder for bilingual people. A Boston University researcher thinks AI can help clinicians make smarter decisions on where to start.
Aphasia is a neurological disorder caused by damage to the brain areas that control language, usually in the left hemisphere. It impairs a person's ability to process language, disrupting their capacity to speak, write, or understand spoken and written words.
- Aphasia affects communication, not intellect.
- It manifests in two ways:
- Expressive (Broca's) aphasia: The person knows what they want to say, but struggles to produce the words or string sentences together.
- Receptive (Wernicke's) aphasia: The person can speak fluently, but the words may not make sense, and has great difficulty understanding others' speech.
Why it matters
Therapy for bilingual aphasia is difficult. Clinicians choose a treatment language based on the patient's preference, or default to English if they don't speak the patient's other language. Neither approach is based on which language would provide the best recovery. A new AI model called BiLex aims to change that.
How it works
- BiLex builds a "digital twin" of each patient's language system, trained on their language history and post-stroke impairment.
- Researchers simulate brain damage and test therapies on the "twin" before treating the patient.
- The model has separate language subsystems but shared meaning representations. This mirrors how the brain organizes multilingual knowledge.
- BiLex explains why language breaks down.
The context
Strokes disrupt the language-switching network, impairing all languages at once. This makes the question of which language to treat first genuinely consequential and hard.
Yes, but
The trial results are complicated.
- At the group level, patients who followed BiLex's recommended therapy language showed no statistically significant improvement over those who didn't.
- The model's real value emerged when researchers broke patients into smaller cohorts by language history and aphasia severity.
- BiLex captured individual differences in language profiles within subgroups.
A closer look
The model isn't ready for broad clinical decisions, but it reads individual patients well, particularly those with severe aphasia or distinct language histories. That's the population where clinicians are flying blind.
The bottom line
BiLex is a proof of concept, not a clinical protocol, and requires further development before clinical application. But the underlying idea adds a new tool for bilingual aphasia care: that a computational model can simulate a person's brain and predict recovery before therapy.
For the 2 million Americans with aphasia, most unaware of the condition’s name, personalized treatment planning has been more hope than reality. This is a step toward changing that.
Concerned about a speech problem?
A screening with a speech-language pathologist will help you understand:
- Your current speech challenge
- Ways to prevent communication barriers
- Staying engaged with loved ones
- Maintaining your quality of life
★ Call 708-599-9500 to schedule your free screening.
★ For facts about speech-language pathology and treatment options, grab your copy of The Consumer's Guide to Speech-Language Pathology.
★ Sign up for our newsletter for the latest on Hearing aids, dementia triggered by hearing loss, pediatric speech and hearing, speech-language therapies, Parkinson's Voice therapies, and occupational-hearing conservation. We publish our newsletter eight times a year.
Don't let speech challenges spoil your enjoyment of life.
Crest Hill, IL - 630-633-5060 | Palos Hills, IL - 708-599-9500
