Quantifying spatially-specific impact of white matter hyperintensity on cortical atrophy in aging
Poster No:
1184
Submission Type:
Abstract Submission
Authors:
Ami Tsuchida1, Iana Astafeva1, Victor Nozais2, Philippe Boutinaud2, Stéphanie Debette1, Marc Joliot1
Institutions:
1Bordeaux University, Bordeaux, Gironde, 2Fealinx, Lyon, Rhône
First Author:
Co-Author(s):
Introduction:
Covert cerebral small vessel disease (cSVD) is common among neurologically asymptomatic older adults, and is a leading cause of stroke and dementia (1). Among the established markers of cSVD, white matter hyperintensity (WMH) is one of the most frequently observed. Although less specific, brain atrophy is also considered as the recurring pathology in cSVD, and it has been hypothesized that it may occur as a secondary degeneration through the damaged white matter tracts (2). However, the direct evidence for the effects of WMH on atrophy in the connected grey matter (GM) regions is lacking. Here we present a preliminary analysis from 800 UK Biobank subjects spanning the age range of 40 to 80 years, testing the regionally-specific effects of WMH on the GM volumes by deriving subject-specific disconnectome maps (3) from the WMH masks.
Methods:
Two hundred UK Biobank subjects (50% female) were randomly selected from each decade of age-range from 40 to 80 years, based on the availability of both WMH lesion map and FreeSurfer-derived grey matter volumetric phenotypes computed by the UK Biobank neuroimaging pipeline (4). Disconnectome maps were computed using the individual WMH mask as a lesion mask in QuickDisco tool in Functionnectome package (5). The mean disconnectivity scores were extracted for each cortical region of DKT atlas (31 regions per hemisphere) using subject-specific FreeSurfer-derived parcellations (Fig. 1). For each region, we tested the impact of WMH-derived disconnectivity scores on the GM volume, in a linear model with GM volume as a dependent variable and age and disconnectivity score as main predictors, and sex and total intracranial volume as additional covariates. We also replaced the disconnectivity score with the overall WMH load to gauge whether disconnectivity derived from WMH rather than the overall load per se had stronger association with regional GM volume variations across subjects.
Results:
Fig. 1 shows the age-related variations in the overall WMH volumes across the sampled population, and the examples of WMH and disconnectome maps derived from them in subjects with extensive or limited WMH. Fig. 2A shows the median disconnectivity scores across DKT atlas regions, indicating a higher disconnection in frontal region, likely as a result of WMH caps commonly found in anterior periventricular regions. The analysis of relative effects of age and WMH disconnectivity on regional GM volume revealed a wide-spread negative effect of age on cross-sectional GM volume variations (Fig. 2B). In contrast, the negative effects of WMH disconnection were limited to a few cortical regions that were somewhat right-lateralized after Bonferroni correction for the 62 regions tested (Fig. 2C). There were paradoxical positive effects on bilateral pericalcarine region, which may indicate erroneous over-segmentation of this cortical region in the proximity of posterior horn WMH. Importantly, repeating the same analysis by replacing the region-specific disconnectivity score with the global WMH load failed to show any significant impact of WMH on GM cortical volume after multiple comparison corrections.
·Fig. 1
·Fig. 2
Conclusions:
Our findings revealed a relatively circumscribed association of WMH-derived disconnectivity scores to the individual variability in the regional GM volumes and wide-spread age effects, suggesting that age-related GM atrophy may be largely independent of WMH-derived disconnection. Nonetheless, disconnectivity scores showed more robust associations with regional GM volumes than global WMH load. Future work is needed to clarify relative contributions of cortical atrophy and WMH disconnectivity to the age-related cognitive decline.
Lifespan Development:
Aging 1
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 2
Segmentation and Parcellation
Keywords:
Aging
STRUCTURAL MRI
Other - White matter hyperintensity; small vessel disease; disconnectome
1|2Indicates the priority used for review
Provide references using author date format
(2) Duering, M. et al. (2023)., 'Neuroimaging standards for research into small vessel disease-advances since 2013', Lancet Neurol, vol. 22, pp. 602–618
(3) Foulon, C. et al. (2018), 'Advanced lesion symptom mapping analyses and implementation as BCBtoolkit', Gigascience, vol. 7, pp. 1–17
(4) Alfaro-Almagro, F. et al. (2018), 'Image processing and Quality Control for the first 10,000 brain imaging datasets from UK Biobank', Neuroimage, vol. 166, pp. 400–424
(5) Nozais, V. et al. (2023), 'Atlasing white matter and grey matter joint contributions to resting-state networks in the human brain', Commun. Biol. vol. 6, pp. 726