The Chatbot program ChatGPT has drawn attention for its ability to create humanlike written responses for some of the most creative queries. Showcasing this, it might one day be able to help doctor’s diagnose Alzheimer’s in its early stages.
A research undertaken at School of Biomedical Engineering, Drexel University recently determined that OpenAI’s GPT-3 program can detect signs from spontaneous speech that are 80% correct in predicting the early stages of dementia.
The Drexel study published in the journal PLOS Digital Health is the latest in a series of efforts to demonstrate the effectiveness of natural language processing programs for early detection of Alzheimer’s. The study leverages current research to suggest that language impairment can be an early indicator of neurodegenerative disorders.
Meanwhile, presently, the current practice to diagnose Alzheimer’s involves a medical history review and long set of physical and neurological examinations and tests. Whilst there is no cure for the disease, early detection can give more alternatives for therapeutics and support. Language impairment is a symptom in 60-80% dementia patients, and for this reason researchers have been focusing on programs that can make out on subtle clues, such as hesitation, grammar and punctuation mistakes and forgetting the meaning of words – as a quick test to indicate if it is required for patients to undergo full examination.
In fact, ongoing research suggests that cognitive effects of Alzheimer’s can manifest themselves in language production, stated a coauthor of the study. Clinically, acoustic features, such as articulation and vocal quality, pausing are commonly used tests for early detection of Alzheimer’s besides cognition tests. But researchers believe improvement of natural language processing programs provide an alternate route for early detection of Alzheimer’s.