WMH are not a homogeneous entity but comprise biologically distinct lesion subtypes with differential neurobiological and clinical significance, suggesting that lesion composition may offer a more informative framework than global WMH burden for understanding cerebrovascular contributions to aging and neurodegeneration.
Key Findings
Results
Three distinct lesion-level WMH subtypes (L1, L2, L3) were identified through unsupervised clustering of longitudinal lesion-wise changes, frequently coexisting within the same individual.
The study analyzed 3224 MRI scans from 403 participants spanning cognitively normal aging, mild cognitive impairment, Alzheimer, and Parkinson disease.
A total of 2107 WMH lesions were identified and characterized using baseline and 2-year follow-up imaging.
Imaging modalities included structural, diffusion, and resting-state MRI.
Clustering robustness was supported by sensitivity analyses excluding anatomical location and by external validation in an independent cohort.
Results
L1 lesions were the most prevalent subtype, predominating in cognitively normal individuals and exhibiting relatively stable trajectories without association with brain atrophy.
L1 lesions accounted for 48.1% of all identified lesions.
This subtype was the most common among cognitively normal participants.
L1 lesions did not show significant associations with brain atrophy.
L1 lesions were characterized by stable longitudinal trajectories compared to the other subtypes.
Results
L2 lesions represented an unstable subtype associated with weight gain, suggesting metabolic vulnerability.
L2 lesions were the least frequent subtype, comprising 11.3% of all lesions.
L2 lesions were significantly associated with weight gain (OR 1.33, 95% CI 1.22–1.45; pFDR ≤ 0.001).
This subtype was characterized as 'unstable' based on its longitudinal trajectory.
The association with weight gain pointed to a metabolic rather than primarily vascular or neurodegenerative mechanism.
Results
L3 lesions represented an unstable subtype significantly associated with brain atrophy, older age, and vascular risk reflected by pulse pressure changes.
L3 lesions accounted for 40.6% of all identified lesions.
L3 lesions were associated with brain atrophy (β = -0.11, 95% CI -0.16 to -0.05; pFDR < 0.001).
L3 lesions were associated with older age (OR 1.15, 95% CI 1.07–1.23; pFDR < 0.001).
L3 lesions were associated with vascular risk reflected by pulse pressure changes (OR 1.09, 95% CI 1.03–1.15; pFDR = 0.006).
External validation in an independent cohort reproduced the main atrophy-related findings for this subtype.
Results
Global WMH burden was no longer associated with brain atrophy after accounting for L3 lesion burden.
This finding suggests that the well-established relationship between global WMH burden and brain atrophy is driven specifically by L3 lesions.
Multivariable models with false discovery rate correction were used to assess these associations.
This result implies that global WMH burden measures may obscure biologically meaningful heterogeneity among lesion subtypes.
The finding supports lesion composition as a more informative framework than global burden measures.
Results
WMH subtypes were identified beyond anatomic location, with sensitivity analyses confirming clustering robustness even when spatial location was excluded.
Sensitivity analyses explicitly excluded anatomical location as a clustering feature to test whether subtypes reflected biology beyond spatial distribution.
External validation in an independent cohort reproduced the main atrophy-related findings.
The study design used longitudinal lesion-wise changes rather than cross-sectional features as the basis for subtype derivation.
All associations were assessed using multivariable models with false discovery rate (FDR) correction.
What This Means
This research suggests that white matter hyperintensities (WMHs) — bright spots visible on brain MRI scans that are commonly used as markers of cerebrovascular disease — are not all the same. Rather than being a single uniform type of lesion, WMHs appear to fall into at least three biologically distinct categories, each with different characteristics and health implications. The researchers tracked over 2,000 individual lesions across 403 participants over two years, using multiple types of brain imaging to understand how each lesion changed over time and what factors it was associated with.
The three subtypes they found behaved very differently. The most common type (about half of all lesions) was relatively stable and was not linked to brain shrinkage or serious risk factors. A smaller group of lesions (about 11%) appeared to be linked to weight gain, hinting at a metabolic connection. The third group (about 41%) was the most clinically concerning: these lesions were unstable, associated with brain atrophy, older age, and higher pulse pressure — a marker of vascular stiffness. Critically, once the researchers accounted for this third subtype specifically, the previously observed link between total WMH burden and brain atrophy disappeared, suggesting that the harmful effects of WMHs are concentrated in this particular subtype rather than spread evenly across all lesions.
This research suggests that treating all white matter lesions as equivalent — as is done when clinicians or researchers simply count total WMH volume — may obscure important biological differences. Understanding which type of lesions a person has, rather than just how many they have, could eventually lead to better risk prediction and more targeted approaches to preventing brain damage associated with aging and neurological diseases such as Alzheimer's and Parkinson's.
Gonzalez-Gomez R, Tagliazuchi E, Campo C, Medel V, Bennett D, Iturria-Medina Y. (2026). Lesion-Level Subtypes of White Matter Hyperintensity Evolution Beyond Spatial Location.. Neurology. https://doi.org/10.1212/WNL.0000000000218472