Cortical integrity changes with cognition, plasma biomarkers and APOE4 status across the adult healthy lifespan

TL;DR

This study explores how brain structure changes with age, cognition, and biomarkers in healthy adults, showing that microstructural changes appear earlier than macrostructural ones.

Contribution

The study introduces new insights into cortical microstructural changes and their relation to cognitive decline and biomarkers in healthy aging.

Findings

cMD shows more extensive age-related changes in younger adults compared to CTh in older adults.

cMD is positively linked to cognitive scores and neurodegeneration biomarkers in prefrontal regions.

cMD and CTh show opposite correlations with inflammation biomarkers and differentiate brain regions.

Abstract

The relationship between cognitive age‐related changes and cortical macrostructural properties (i.e., cortical thickness [CTh]) has been extensively studied. However, less is known about the relation with microstructural characteristics (i.e., cortical mean diffusivity [cMD]) even though these are sensitive to preclinical phases of Alzheimer's disease (AD). Furthermore, the relationship between macro‐ and microstructural measures with cognition, neurodegeneration and inflammation plasma biomarkers among healthy individuals has not been reported.

A total of 964 adults (age: 40‐82 years; 52% females) with normal neuropsychological profile and available structural and diffusion MRI data were included. FreeSurfer was used to obtain CTh maps. For cMD processing we used a homemade surface‐based approach after preprocessing the diffusion images with FSL. Surface maps were smoothed (FWHM=15) and the group vertex‐wise statistical maps were considered significant at p <0.05. The Preclinical Alzheimer's Cognitive Composite (PACC) was used as the cognition measure. Plasma concentrations of phosphorylated tau (p‐Tau181), neurofilament light (NFL) and C‐reactive protein (hsCRP) were analyzed, together with APOE4 status.

We identified a significant negative association between both cortical measures and age (Figure 1). cMD associations were more extensive at earlier ages (under 50 years), while CTh associations were greater at older ages (above 50 years) (Figure 2). cMD was positively correlated with PACC scores and both pTau‐181 and NFL concentrations in prefrontal regions, while the association was negative and more widespread for hsCRP (Figure 3). CTh results revealed smaller positive and negative clusters only for PACC and NFL, respectively (Figure 3). Correlating cMD with CTh differentiated somatosensory and associative areas with negative and positive correlations, respectively (Figure 3). This pattern was conserved after covarying for sex, age and biomarker data, and across age and risk (APOE4 carrier) subsamples.

We show that cMD, which might reflect neuronal density loss accompanying or even preceding atrophy, is able to capture microstructural cortical changes occurring during natural aging before CTh alterations. Indeed, it seems more sensitive to age‐related cognitive decline. Furthermore, our results suggest a pattern relating the two metrics conserved across different aging trajectories and opposite trends of the two cortical measures in relation to neurodegenerative biomarkers levels.

Figures