Cortical mean diffusivity detects early age-related changes and associates with cognition and plasma biomarkers

The relationship between age-related cognitive changes and cortical macrostructural properties [i.e. cortical thickness (CTh)] has been extensively studied. However, less is known about the relationship with cortical microstructural characteristics [i.e. cortical mean diffusivity (cMD)] even though these are sensitive to preclinical phases of Alzheimer's disease. We studied a sample of 964 cognitively healthy adults (age: 40-82 years; 52% females) with available structural and diffusion MRI data. The preclinical Alzheimer's cognitive composite was used as the cognition measure, and plasma concentrations of neurodegenerative-related (i.e. phosphorylated tau 181 and neurofilament light) and inflammatory (i.e. high-sensitivity C-reactive protein) biomarkers were assessed, together with apolipoprotein ɛ4 status. Neuroimaging data was preprocessed using FreeSurfer and FSL, and a homemade surface-based approach was used to obtain cMD maps. A two-class generalized linear model was used as the main statistical analysis. We identified a significant negative association between both cortical measures (cMD and CTh) and age. cMD associations were more extensive at earlier ages (<50 years), while CTh associations were greater at older ages (>50 years). cMD was positively correlated with cognition and with both neurodegenerative-related biomarkers in prefrontal regions, while the association was negative and more widespread for the inflammatory biomarker. CTh was positively correlated with cognition in more restricted areas than cMD and only negatively correlated with neurofilament light. Also, cMD presented lower levels in apolipoprotein ɛ4 carriers compared to non-carriers, while no results were found for CTh. Correlating cMD with CTh resulted in a regional pattern of negative and positive correlations, differencing somatosensory and associative areas, respectively. Altogether, we show that cMD can capture microstructural cortical changes occurring across adulthood into older age before CTh alterations. Indeed, it seems more sensitive to age-related cognitive decline and pathological and inflammatory processes related to risk profiles, showing an opposite trend to CTh in relation to neurodegenerative biomarker levels. Furthermore, our results suggest a pattern relating the two cortical metrics perhaps reflecting a cortical organization pattern.