Population-referenced multiregional brain-volume profiles associated with performance and concussion history in athletes: an exploratory study.

INTRODUCTION

Elite athletes may exhibit structural brain profiles shaped by long-term training, selection into elite sport, exposure to repetitive head impacts, or combinations of these factors. Interpreting brain structure in this population is challenging because general-population reference values may not distinguish adaptive athlete-associated variation from injury-related or clinically meaningful change. In this exploratory cross-sectional study, we used age- and sex-referenced regional brain-volume percentiles as a population-referenced framework to examine whether multiregional structural profiles in elite athletes were associated with athletic performance, sensorimotor function, and reported concussion history.

METHODS

Fifty-one active and retired elite athletes underwent 3T structural MRI and robotic sensorimotor testing (KINARM). T1-weighted scans were segmented using BrainKey (brainkey.ai) to obtain age- and sex-referenced regional percentiles across 18 brain regions. Principal component analysis (PCA) was applied directly to the regional percentile measures, and Bayesian models assessed associations with ordered performance ratings, KINARM components, and reported concussion history.

RESULTS

Five principal components explained 71.4% of the variance. Brain PC1 was positively associated with higher performance-group membership [β = 0.53, 95% CrI (0.04, 1.02), posterior probability of direction = 0.983, n = 51]. Brain PC3 was positively associated with KINARM PC3 [β = 0.57, 95% CrI (0.01, 1.13), posterior probability of direction = 0.978, n = 44]. However, both components showed limited bootstrap stability, making this result highly exploratory. Reported concussion history was available for 50 participants and showed suggestive directional evidence for lower Brain PC2 [OR = 0.61, 95% CrI (0.36, 1.04), posterior probability of direction = 0.97] and higher Brain PC1 [OR = 1.57, 95% CrI (0.92, 2.71), posterior probability of direction = 0.95].

DISCUSSION: These exploratory findings indicate that population-referenced multiregional brain-volume profiles may be related to athletic performance, sensorimotor function, and reported concussion history. The results do not establish pathological or beneficial structural brain change and do not construct or validate an athlete-specific normative model. Instead, they identify candidate multiregional profiles for evaluation in larger, longitudinal athlete cohorts with matched non-athlete comparison groups.