OVERVIEWAiming to predict changes in individual brain health, this functional magnetic resonance imaging (fMRI) study uses a machine learning model trained on brain imaging data. This new research highlights the importance of brain health, emphasizing the holistic nature of cognitive health, well-being and connectedness. Participants from ages 21–65 completed sensorimotor tasks while undergoing two brain scans six months apart. These fMRI scans measured hemodynamic response functions (HRFs) using both traditional and new metrics, with the machine learning model accurately predicting significant improvements in cognitive brain health with 90% accuracy, and shedding new light on three crucial HRF factors: changes in amplitude and dispersion, and the participant's HRF shape at baseline.Findings suggest the uncharted potential of brain imaging and innovative machine learning metrics to effectively track brain health in healthy individuals. Researchers also address the need for a more consistent definition of brain health and increased attention to the concept and goal of "precision brain health."
Fig. 4. Probability surface for the final neural network model predicting change in amplitude (x-axis).