Predicting Alzheimer’s disease with GPT-3

Predicting Alzheimer’s disease with GPT-3 with 80% accuracy

Although there is currently no cure for Alzheimer’s disease, making a diagnosis earlier in the disease can still provide patients with more treatment and options.

The current diagnosis of early stage Alzheimer’s disease requires lengthy assessment tests of the patient’s body and mind.

A recent study published in PLOS Digital Health by the School of Bioengineering at Drexel University (USA) shows that early-stage Alzheimer’s disease can be diagnosed using GPT-3, a text-generation tool from OpenAI.

60% to 80% of patients with Alzheimer’s disease have symptoms of speech impairment.

Researchers have been trying to diagnose Alzheimer’s disease by capturing subtle clues such as hesitations, speech disorders and mispronunciations in patients’ speech.

The researchers analyzed the patients’ voice data with the GPT-3 and had the GPT-3 make judgments based on the Mini-Mental State Examination (MMSE).

The researchers found that the GPT-3 predicted MMSE scores based on voice 20% higher than traditional methods and was up to 80% more accurate in predicting early Alzheimer’s disease.

Leave a Comment