AI May Make Chairside Diagnosis of Cognitive Impairment a Reality
New research suggests AI can analyze subtle changes in speech during routine patient-provider interactions to identify early signs of cognitive impairment, potentially giving health and dental professionals a new tool for early detection
While the debate over the use of artificial intelligence (AI) wages on (will humans be wiped out by rogue superintelligence?), researchers have found a strictly positive function for current AI models. According to a study published in JAMA Neurology, scientists were able to train AI to detect the first signs of cognitive impairment in adults age 55 and older during mundane interactions between patients and physicians.
The AI models were fed 30-second patient-provider interactions recorded during health visits for 966 patients. AI was able to use acoustic features, including pitch, timing, and variability, to detect cognitive impairment among the patient cohort, which had no recorded cognitive impairments prior to the study. The AI models reached a sensitivity of 68%.
Cognitive impairments can be more effectively addressed when caught early; however, due to lack of time and training, many are not diagnosed until the impairments are much more advanced. Providing a tool to healthcare and dental providers that can quickly and easily detect such impairments would be a game changer.Click here to read the study.
