White matter hyperintensities, the bright patches that show up on routine brain scans in later life, are not uniform inside, according to a quantitative MRI study published in the journal Brain on 6 June 2026. Aurelie Bussy and colleagues at the University of Bern and Lausanne University Hospital scanned 363 community-dwelling adults from the BrainLaus cohort (mean age 55.5 years, range 19.8 to 89.4, 48.8 per cent male) and reported layer-specific profiles showing what the authors describe as a centrifugal gradient: myelin loss and oedema at the lesion core, and axonal alteration in the surrounding tissue.

The team combined quantitative relaxometry (MTsat, R1 and R2*, sensitive to myelin and iron) with diffusion measures (fractional anisotropy, mean diffusivity, NODDI and the g-ratio, sensitive to axons and to water outside cells). Lesions were segmented automatically on FLAIR images, and microstructure was sampled across lobes, white matter compartments, and geodesic layers running from the lesion core out into normal-appearing white matter. The overall pattern was demyelination, axonal loss and extracellular fluid accumulation, most pronounced in periventricular regions. Using partial least squares, the authors also found these signatures were associated with cardiovascular risk factors including higher blood pressure, bioimpedance measures and lower haemoglobin. Those associations are statistical, measured at one moment, and do not establish that any risk factor caused the tissue changes.

The finding matters because lesion volume is the number usually carried forward, into risk scores and into trials, as though every millilitre of hyperintensity were the same tissue. The authors' conclusion is that hyperintensities index the endpoint of a broader, spatially structured injury process that extends into white matter still classed as normal on a standard scan. White matter hyperintensities are the imaging hallmark of cerebral small vessel disease, the vascular process that contributes to vascular dementia and to the mixed pathology found in many people diagnosed with Alzheimer's disease.

The limits are substantial and the authors set them out. No brain tissue was examined in these participants, so every claim about myelin, axons and fluid rests on MRI measures that stand in for those things rather than measuring them directly. The design is cross-sectional, a single snapshot, so it cannot show whether lesions grow outward from a damaged core or whether the surrounding change comes first. An earlier version of the work, posted as a preprint on bioRxiv on 2 October 2025, used a smaller and older subsample of 202 participants; the peer-reviewed paper reports the larger group. Volunteers in a population cohort also differ from patients seen in memory clinics.