Patterns of cortical degeneration in an elderly cohort with cerebral small vessel disease

Abstract

Emerging noninvasive neuroimaging techniques allow for the morphometric analysis of patterns of gray and white matter degeneration in vivo, which may help explain and predict the occurrence of cognitive impairment and Alzheimer's disease. A single center prospective follow-up study (Radboud University Nijmegen Diffusion tensor and Magnetic resonance imaging Cohort study (RUN DMC)) was performed involving 503 nondemented elderly individuals (50-85 years) with a history of symptomatic cerebral small vessel disease (SVD). Age was associated with a global reduction in cortical thickness, and this relationship was strongest for ventrolateral prefrontal cortex, auditory cortex, Wernicke's area, superior temporal lobe, and primary visual cortex. Right and left hemispheres differed in the thickness of language-related areas. White matter (WM) lesions were generally negatively correlated with cortical thickness, primarily in individuals over the age of 60, with the notable exception of Brodmann areas 4 and 5, which were positively correlated in age groups 50-60 and 60-70, respectively. The observed pattern of age-related decline may explain problems in memory and executive functions, which are already well documented in individuals with SVD. The additional gray matter loss affecting visual and auditory cortex, and specifically the head region of primary motor cortex, may indicate morphological correlates of impaired sensory and motor functions. The paradoxical positive relationship between WM lesion volume and cortical thickness in some areas may reflect early compensatory hypertrophy. This study raises a further interest in the mechanisms underlying cerebral gray and white matter degeneration in association with SVD, which will require further investigation with diffusion weighted and longitudinal MR studies.

Figure 1

Cortical grey matter regions of interest (GM ROIs). A: Parcellation of the cortical surface into Brodmann areas. Grey areas are not analyzed in the present study, and include the medial wall (MW), which is artificially created by the surface extraction process, and the insula region. B: Boxplots of Thickness for each Brodmann area ROI, showing the distributions for both hemispheres. ROIs are sorted by increasing mean Thickness. Filled boxes represent quartiles 1 to 3, capped lines indicate the range of regular values, circles indicate the mean, and horizontal lines indicate the median.

Figure 2

Three‐dimensional renderings of vertex‐wise values mapped onto the average surface and scatterplots of individual Brodmann area ROIs. A: Spatial distribution of vertex‐wise mean. Thickness in mm. B: Spatial distribution of the slope for Age in each vertex‐wise linear model of the form Thickness ∼ Age + Sex + Age × Sex. C: Spatial distribution of the slope for Sex (Male–Female) in the model Thickness ∼ Age + Sex; a positive (red) value corresponds to Male > Female. D: Scatterplots of Age versus mean Thickness for five Brodmann areas.

Figure 3

A: Bar plots showing the slope, standardized to the first‐order population statistics, for total WMLV in individual GM ROI‐wise linear regression models of the form Thickness ∼ log(WMLV). Analyses were performed over three age groups: 50–60, 60–70, and 70+. Model significance, accounting for FDR, is represented with asterisks (* indicates P < 0.05, q < 0.05; ** indicates P < 0.05, q < 0.01). For each lobe, BAs are sorted by the slope values from the 60–70 group. B: Selected scatterplots with linear regression lines for each age group, showing mean ROI‐wise cortical thickness versus log(WMLV) for BA4 and BA5 (positive), and BA10 (negative). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]