Artificial Intelligence Models Identify Alzheimer's Cognitive Decline

Massachusetts Institute of Technology (MIT) researchers have developed artificial intelligence tools to determine whether individuals at high risk for Alzheimer's will suffer cognitive decline, reports Health IT Analytics . The algorithm predicted cognition test scores up to two years in the future, and the resulting models could help with selecting candidate drugs and participants for clinical trials, as well as help persons and their loved ones prepare for this decline. The MIT team initially trained a population model on an entire dataset that included clinically significant cognitive test scores and other biometric data from people with Alzheimer's and healthy individuals between biannual checkups. A second model is more personalized for each subject, and continuously updates risk scores based on newly posted data. To optimize the models' results, the team developed a "metalearning" framework that automatically selects the best model for any given participant at any given time. The framework cut the error rate for future predictions by 50 percent. Trials showed the models can produce accurate forecasts six, 12, 18, and 24 months in advance.