Cognitive intraindividual variability and white matter integrity in aging

Abstract

The intraindividual variability (IIV) of cognitive performance has been shown to increase with aging. While brain research has generally focused on mean performance, little is known about neural correlates of cognitive IIV. Nevertheless, some studies suggest that IIV relates more strongly than mean level of performance to the quality of white matter (WM). Our study aims to explore the relation between WM integrity and cognitive IIV by combining functional (fMRI) and structural (diffusion tensor imaging, DTI) imaging. Twelve young adults (aged 18-30 years) and thirteen older adults (61-82 years) underwent a battery of neuropsychological tasks, along with fMRI and DTI imaging. Their behavioral data were analyzed and correlated with the imaging data at WM regions of interest defined on the basis of (1) the fMRI-activated areas and (2) the Johns Hopkins University (JHU) WM tractography atlas. For both methods, fractional anisotropy, along with the mean, radial, and axial diffusivity parameters, was computed. In accord with previous studies, our results showed that the DTI parameters were more related to IIV than to mean performance. Results also indicated that age differences in the DTI parameters were more pronounced in the regions activated primarily by young adults during a choice reaction-time task than in those also activated in older adults.

Figure 1

Radiological convention—right of the subject on the left of the figure. (a) Young > old: map of regions significantly more activated (in red) by the young population than by the older population in the RT task (CrossSquare). The right prefrontal (superior frontal gyrus) region is activated in the younger adults, as well as the angular gyrus (right more than left). (b) Old > young: map of regions significantly more activated (in red) by the old population than by the younger population in the RT task. There is significantly more right insular activation in the older population, as well as some larger bifrontal, supplementary motor areas and parietal networks (perirolandic).

Figure 2

Conjunction of young and old adult activation: regions activated by both groups are highlighted in red. They comprise the right cerebellum, the right inferior temporal gyrus, the right striatum, the right occipital cortex, bifrontal rolandic cortex (hand area), and bilateral prefrontal cortex. Neurological convention—right of the subject on the axial image on the right of the figure.

Figure 3

Creation of the white matter mask from the fMRI-activated areas: the yellow regions are regions activated during the contrast analysis of the young population (young only). The activated areas are then enlarged to the adjacent brain by a radius of 5 mm (blue regions). Then a mask is applied to keep the white matter only (not represented on the figure).

Figure 4

Using the JHU atlas, diffusivity measures were computed for both groups (young and old adults). The upper row represents the regions where the FA values were higher in the younger adults (forceps major and minor, bilateral inferior fronto-occipital fasciculus, and bilateral inferior longitudinal fasciculus). Mean values across areas: .478 (SD = .015) versus .448 (.012). The lower row represents the regions where the MD values were higher in the older population (bilateral anterior thalamic radiation, forceps minor, left IFO, and bilateral superior longitudinal fasciculus). Mean values across regions: .00075 (.00002) versus .00084 (.00006).

Figure 5

Scatterplots of iSD in the complex processing speed task and DTI characteristics (FA, MD, RD, and AD). For illustration, only ROIs showing significant correlations with IIV have been considered. IIV in processing speed was negatively related to fractional anisotropy and positively related to diffusivity (MD, AD, and RD).