Neuroimaging predictors of cognitive impairment in confluent white matter lesion: volumetric analyses of 99 brain regions.

BACKGROUND: Although confluent white matter lesion (WML) is associated with cognitive impairment, the mechanism explaining this association is controversial. We aimed to investigate comprehensively the MRI predictors of cognitive impairment in confluent WML. METHODS: Among 45 lacunar stroke patients who had confluent WML, we evaluated the association of executive function [Mattis Dementia Rating Scale - Initiation/Perseveration subscale (MDRS I/P)] and global cognition [Mini-Mental State Examination (MMSE)] with the volume of WML, measures of lacunes and microbleeds, and the volumes of 99 other specific brain regions. RESULTS: Regression analyses showed that WML volume predicted performance on the MDRS I/P (beta = -0.34, p = 0.016) independent of age. Volumes of cortical gray matter (cGM; beta = 0.41, p = 0.003), the lateral fronto-orbital gyrus (beta = 0.38, p = 0.01), superior frontal gyrus (beta = 0.29, p = 0.04), lateral ventricle (beta = -0.30, p = 0.04), and posterior limb of the internal capsule (beta = 0.43, p = 0.002) predicted MDRS I/P performance independent of WML volume. Volumes of cGM, and the lateral fronto-orbital gyrus predicted MMSE performance as well. CONCLUSION: Atrophy along the frontosubcortical pathways and cGM predict cognition in confluent WML independent of WML volume.