Game-Based Cognitive Aging Assessment: Toward a Digital Biomarker of Cognitive Health

Area of Science:

• Neuroscience
• Gerontology
• Computational Psychology

Background:

• Cognitive performance naturally declines with age, impacting life outcomes.
• Cognitive aging is an underutilized biomarker for assessing health and development.
• Digital biomarkers offer a novel approach to cognitive assessment.

Purpose of the Study:

• To establish the feasibility and value of game-based digital biomarkers for cognitive aging.
• To utilize a home-based cognitive assessment game for data collection.
• To develop and validate machine learning models for predicting cognitive aging.

Main Methods:

• 871 participants (ages 18-75) played the Tunnel Runner cognitive game.
• 17 game-based cognitive metrics were used to train machine learning models to predict chronological age.
• Nested cross-validation was employed to assess out-of-sample predictive accuracy.

Main Results:

• The best machine learning model predicted age with a 6.97-year mean absolute error (correlation=0.626).
• No predictive bias was found for gender or gaming identity.
• Cognitive aging scores correlated with ADHD (reduced aging) and gamer identity (negative association).

Conclusions:

• Game-based assessments can generate accessible digital biomarkers of cognitive aging.
• This approach allows for scalable, low-burden cognitive aging assessment.
• Potential applications include early cognitive decline detection, tracking, and intervention evaluation.