Connecting the dots: microstructural properties of white matter hyperintensities predict longitudinal cognitive changes in ageing.

Abstract

This study investigates the relationship between white matter hyperintensities (WMHs) and longitudinal cognitive decline in older adults. Using data from The Irish Longitudinal Study on Ageing (TILDA), we examined WMH characteristics, including volume, location, and microstructural integrity, in a community-dwelling population of 497 individuals over a six-year period. WMHs were categorised into phenotypes based on their size, fractional anisotropy (FA), and mean diffusivity (MD), with subtypes for periventricular and deep white matter lesions. We hypothesised that larger, microstructurally compromised lesions would be associated with accelerated cognitive decline. We isolated 11,933 WMHs, with an average of 24 WMHs per individual. Of these lesions, 6,056 (51%) were classified as Low Volume - High FA, 3193 (27%) were classified as Low Volume - Low FA and 2684 (22%) were classified as High Volume, Low FA. Our findings demonstrate that high-volume, low FA deep (p = 0.05) and periventricular (p = 0.004) lesions were significantly linked to cognitive decline (X = 12.9, p = 0.004), whereas small periventricular lesions with near normal microstructural properties do not predict cognitive decline. These results suggest that distinct WMH phenotypes may serve as markers for differential risks of cognitive impairment, providing potential targets for early intervention in at-risk populations.