Understanding why tau pathology targets specific brain regions — and spares others — is one of the central puzzles in neurodegenerative disease. For the roughly 2,500 former NFL players and tens of thousands of contact-sport athletes who may carry elevated CTE risk, this question carries direct clinical relevance. New imaging research offers the most mechanistically detailed answer to date, pointing to two structural properties that may predetermine where tau takes hold.
The study combined three independent datasets — tau-PET imaging in 157 former professional American football players and 53 controls, cortical myelin water fraction (MWF) maps from 50 healthy individuals, and white matter tract MWF plus functional connectivity data from 100 participants in the Human Connectome Project. Cortical gray matter regions with lower baseline myelination showed significantly higher tau-PET uptake (β = −0.399, p = 0.001). Separately, white matter tracts with lower typical MWF correlated with greater tau covariance across the brain (β = −0.238, p < 0.001), while higher functional connectivity between regions predicted broader tau co-distribution (β = 0.447, p < 0.001). In short, poorly myelinated, highly connected nodes appear to serve as both preferential deposition sites and spreading hubs.
This work meaningfully advances the CTE field by proposing a biological explanation — reduced myelin integrity and elevated network centrality — for tau's non-random topography, rather than treating spread as purely stochastic. The myelin-vulnerability link resonates with emerging evidence in Alzheimer's disease, where regions that myelinate late developmentally also accumulate amyloid and tau earliest. The functional connectivity finding parallels prion-like spreading models demonstrated in Parkinson's and frontotemporal dementia research. Important caveats apply: the football player cohort is living and at-risk, not confirmed CTE cases (definitive diagnosis still requires post-mortem tissue); reference myelination norms were drawn from separate, smaller healthy cohorts rather than matched controls; and causality cannot be established cross-sectionally. Still, the triangulated multi-dataset approach strengthens confidence. This is a genuinely informative mechanistic step, though replication in larger longitudinal cohorts with autopsy confirmation is essential before clinical translation.