BrainHealth

Background: Optimization of brain health is a focal point in medical science, yet data regarding measuring, preserving, and improving lifelong brain health are lacking. This void demands an objective, change-sensitive measure of brain health and proven strategies to strengthen brain performance. The BrainHealth Project addresses these key issues, drawing upon neuroplasticity evidence of persistent modifiability of brain function across the lifespan. This landmark study aims to definitively (i) evaluate and refine holistic change indices of brain health, (ii) evaluate the impact of evidenced-based cognitive strategies and lifestyle interventions on improving/maintaining brain health, and (iii) elucidate the mechanisms associated with brain health gains/losses. Methods: This prospective, longitudinal, interventional, open-label, single-arm clinical trial aims to recruit 100,000 generally healthy adults over 10 years. Assessments, coaching, and training are conducted online through the secure BrainHealth Platform, allowing utilization tracking. The BrainHealth Index (BHI)—a multidimensional assessment—is offered at baseline and every six months. Participants have access to coaching every three months and continual access to self-paced trainings and resources. The primary outcome is the further validation of a novel composite BHI score and its sub-scores: Clarity (cognitive health), Connectedness (social health), and Emotional Balance (well-being). The BHI includes measures sensitive to change, including gains from behavior change and integration of cognitive strategies into daily life. These contrast with traditional assessments focused on detecting cognitive decline or diagnosing pathological conditions. The primary online training, SMART (Strategic Memory Advanced Reasoning Tactics), is a strategy-based program validated with 25+ years of research. SMART promotes improvements in neural health, cognition, well-being, connectedness, and real-life function, previously reported in randomized controlled trials (RCTs). Statistical approaches focus on individual prediction using nonlinear models trained with large samples and on assessing mechanisms influencing gains or losses on brain health metrics. Discussion: This research extends RCT evidence to a longitudinal, epidemiologic approach, leveraging digital health and machine learning tools to deliver a generalizable measure of longitudinal brain health and to achieve precision brain health. By integrating advanced statistical methods and large-scale data, the BrainHealth Project should provide medicine and society with accurate and actionable ways to optimize brain health practices across the lifespan.

Abstract The goal of precision brain health is to accurately predict individuals' longitudinal patterns of brain change. We trained a machine learning model to predict changes in a cognitive index of brain health from neurophysiologic metrics. A total of 48 participants (ages 21–65) completed a sensorimotor task during 2 functional magnetic resonance imaging sessions 6 mo apart. Hemodynamic response functions (HRFs) were parameterized using traditional (amplitude, dispersion, latency) and novel (curvature, canonicality) metrics, serving as inputs to a neural network model that predicted gain on indices of brain health (cognitive factor scores) for each participant. The optimal neural network model successfully predicted substantial gain on the cognitive index of brain health with 90% accuracy (determined by 5-fold cross-validation) from 3 HRF parameters: amplitude change, dispersion change, and similarity to a canonical HRF shape at baseline. For individuals with canonical baseline HRFs, substantial gain in the index is overwhelmingly predicted by decreases in HRF amplitude. For individuals with non-canonical baseline HRFs, substantial gain in the index is predicted by congruent changes in both HRF amplitude and dispersion. Our results illustrate that neuroimaging measures can track cognitive indices in healthy states, and that machine learning approaches using novel metrics take important steps toward precision brain health.