Genetics

This post covers Feng and colleagues' 2026 study in Magnetic Resonance in Medical Sciences, which tested whether a polygenic risk score for multiple sclerosis relates to brain white matter microstructure and cognition in 35,952 neurologically healthy UK Biobank participants. The primary analysis is a null across all 11 cognitive measures. The post converts the one surviving subgroup finding into milliseconds, walks through the single imaging metric that survived out of seven tested in eight tracts, examines what a 3% mediation effect means, credits the authors' unusually careful framing of their own exploratory result, and closes on the vascular and schooling covariates left out of the model.