The relevance of tract‐specific white matter hyperintensity burden to clinical impairment in groups at risk of Alzheimer's disease.

Background: White matter hyperintensities (WMHs) are common in normally functioning older adults. While greater WMH volume may be associated with an increased risk of cognitive dysfunction and dementia, the impact of the location of WMHs is less well understood. We investigated whether WMH burden in specific white matter (WM) tracts is associated with clinical classification and global cognitive scores in at‐risk groups for Alzheimer's disease (AD). Method: Participants from Dementia Prevention Research Clinics, New Zealand, were classified as control, subjective cognitive decline, single‐domain amnestic mild cognitive impairment (MCI), multiple‐domain amnestic MCI, or early AD by a multidisciplinary team (Table 1). Several tract‐WMH "overlap scores" (the percentage of total tract voxels containing WMHs) were calculated for each participant's binarised WMH map (Lesion Segmentation Toolbox, SPM) using seven tract probability templates from a 140‐subject tract atlas (www.megatrackatlas.org). Tract templates were of the superior longitudinal fasciculus (SLF), inferior longitudinal fasciculus (ILF), inferior fronto‐occipital fasciculus (IFOF), uncinate fasciculus (UF), cingulum (CB), fornix, and corpus callosum (CC). Between‐group differences in overlap scores were tested with ANOVAs. Pearson's correlations examined associations between tract‐WMH overlap and global cognitive scores. Result: Figure 1 shows WMH overlap with one tract, the IFOF, in a representative participant. Considering the whole sample, the highest tract‐WMH overlap scores can be seen in the IFOF and CC. Significant between‐group differences were found for the SLF, ILF, IFOF, UF, and CC (p <.05), but not the CB or fornix (Fig. 2), with significant linear trends describing the relationship between overlap scores and clinical classification. Post‐hoc tests confirmed significant pairwise comparisons for overlap scores in all aforementioned tracts except the SLF (Fig. 2). We found Addenbrooke's Cognitive Examination‐III scores negatively correlated with overlap scores in the ILF (r = ‐0.19), IFOF(r = ‐0.22), UF(r = ‐0.22), and CC(r = ‐0.19) (all p <.01), but not the SLF, CB, or fornix (p >.05). Conclusion: Our findings highlight that WMH burden within specific WM tracts is associated with degree of clinical impairment in groups at‐risk of AD. Future research will investigate relationships between overlap scores in specific WM tracts and executive functioning, episodic memory, and processing speed performance. [ABSTRACT FROM AUTHOR]