New study uses AI to predict who may benefit from cognitive training to stave off dementia

Person pointing at computer with brain model in foreground.
Dr. Joseph Gullett. Photo by Louis Brems.

By Jill Pease

A new University of Florida College of Public Health and Health Professions study aims to shed light on which older adults at risk for Alzheimer’s disease may respond to cognitive training programs designed to boost thinking skills that tend to decline with aging and disease.

Led by Joseph Gullett, Ph.D., an assistant professor in the PHHP department of clinical and health psychology who was hired under UF’s artificial intelligence initiative, the study will use a combination of brain imaging and artificial intelligence tools to establish the effectiveness of a take-home, 12-week cognitive training program. The project is supported by a five-year K23 research career development grant from the National Institute on Aging totaling more than $800,000.

The study focuses on patients with a diagnosis of amnestic mild cognitive impairment. According to the Alzheimer’s Association, amnestic mild cognitive impairment may affect patients’ ability to remember important information, such as appointments, conversations or recent events. Some people with amnestic mild cognitive impairment may go on to develop Alzheimer’s disease.

“Knowing that their diagnosis of amnestic mild cognitive impairment places them at risk for developing Alzheimer’s disease, patients often seek out intervention to delay the progression of the disease without knowing whether it will truly provide benefit,” said Gullett, a member of the UF Center for Cognitive Aging and Memory Clinical Translational Research. “This project will use advanced machine learning tools to provide detailed predictions of patients’ likelihood of improving after participating in a widely-supported cognitive training intervention. It will also clarify how an individual patient’s brain and cognitive features contribute to their response to the intervention.”

Patients in the early stages of Alzheimer’s disease each display a unique pattern of altered brain structure and function, as well as thinking and memory skills. These unique characteristics may hold the key for understanding who will respond to cognitive training and who will not.

“This study is designed to use powerful AI tools to identify brain regions responsible when person does not respond to a proven cognitive training intervention,” Gullett said. “In the future, these findings may allow us to use targeted and safe brain stimulation on these specific brain regions to increase the chance that at-risk individuals will respond more strongly to cognitive training.”