Multi-domain Brain Age from Digital Cognitive Metrics as a novel approach for new longevity
Arbizu-Gomez, M.; Sastre-Barrios, C.; Maltseva, E.; M. Corada, J.; Ortea Suarez, C.; Fernandez de Pierola, I.; Lubrini, G.; Perianez, J. A.; Rios-Lago, M.; Cortes, J. M.
Show abstract
BackgroundThe continuous rise in life expectancy introduces a central challenge of new longevity, ensuring that the additional years gained are accompanied by the preservation of cognitive function and quality. MethodsWe propose a modeling framework for multi-domain brain age derived from a repertoire of digital cognitive metrics. The model, based on Ridge regression with Leave-One-Out cross-validation, was trained in a cohort of 394 healthy controls (HC; 307 women and 87 men; mean age 30.0 {+/-} 12.5 years; range 17-64). ResultsThe model achieved a correlation between chronological age and predicted age of r = 0.942 with a mean absolute error of 3.05 years. When applied to three additional clinical cohorts, multiple sclerosis (N = 70), traumatic brain injury (N = 23), and depression (N = 18), the model detected significant accelerated cognitive aging across all conditions, with processing speed emerging as the dominant contributor to accelerated aging, albeit with varying degrees of concentration across pathologies. ConclusionsDigital cognitive metrics provide an accessible, non-invasive, and scalable biomarker for tracking brain aging, with strong potential for informing personalized neuropsychological interventions and for integration into active aging frameworks within the context of modern longevity.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.