Pitfalls in brain age analyses

Area of Science:

• Neuroimaging
• Biomarkers
• Aging Research

Background:

• The brain age gap, comparing predicted brain age to chronological age, is a widely studied metric.
• Previous research indicates the brain age gap is dependent on chronological age.
• Group differences in the brain age gap may reflect age disparities rather than true biological differences.

Purpose of the Study:

• To critically evaluate the methodology of brain age gap analyses.
• To address the limitations of current approaches in quantifying deviations from normal brain aging.
• To propose theoretical advancements for accurately measuring brain age discrepancies.

Main Methods:

• Analysis of the brain age gap as a linear transformation of residuals.
• Examination of the impact of regressing chronological age out of the brain age gap.
• Theoretical assessment of model accuracy statistics (e.g., R-squared) following age correction.

Main Results:

• The brain age gap is inherently dependent on chronological age.
• Regressing age out of the brain age gap artificially inflates model accuracy metrics (R-squared).
• Achieving R-squared values below 0.85 after age correction is highly improbable, irrespective of true model accuracy.

Conclusions:

• Current methods for analyzing the brain age gap are limited by their dependence on chronological age.
• Age correction inflates statistical measures, potentially misrepresenting the true accuracy of brain age prediction models.
• Further theoretical development is required to establish robust methods for quantifying deviation from normal brain aging.