Pointers at Glance
- Artificial Intelligence (AI) is becoming increasingly prevalent in the medical field, potentially revolutionizing how we diagnose and treat disease.
- One of the most exciting applications of AI is the development of AI driving chatGPT to detect early signs of Alzheimers disease.
ChatGPT is a natural language processing (NLP) system capable of understanding and responding to human conversations. It is used to identify subtle changes in a person’s speech and language patterns which may indicate the early onset of Alzheimer’s disease.
By utilizing AI to identify these early signs, it is possible to increase the probability of early diagnosis and treatment, allowing patients to live better and longer lives.
Finding an Early Sign
Currently, the practice for diagnosing Alzheimer’s Disease usually involves a medical history review and a lengthy set of physical and neurological evaluations and tests. However, as there is no cure for this disease, detecting it earlier can support patients with multiple options for therapeutics.
For early detection of Alzheimers Disease, the commonly used tests look at acoustic features like pausing, articulation, and vocal quality, in addition to cognition tests. But, the improvement of NLP programs offers another path to support early disease identification.
A Program That Listens and Learns
GPT-3 has undergone sufficient training to understand the reference and adapt to produce the expected response.
Felix Agbavor, a doctoral researcher in the School and the lead author of the paper, said that Training GPT-3 with a vast dataset of interviews with Alzheimer’s patients would provide it with the information it requires to extract speech patterns. That could then be applied to identify markers in future patients. GPT-3 is considered almost 20% more accurate in predicting the MMSE scores of patients.
After conducting multiple tests, the results demonstrate that the text embedding, generated by GPT-3, can be reliably used to detect individuals with Alzheimer’s Disease from healthy controls and infer the subject’s cognitive testing score based on speech data.
The researchers plan to build a web application that could be used at home or in a doctor’s office as a pre-screening tool.
