Disconnectome Associated with Progressive Ischemic Periventricular White Matter Lesions

Abstract

Periventricular white matter (PVWM) hyperintensities on T2-weighted MRI are ubiquitous in older adults and associated with dementia. Efforts to determine how PVWM lesions impact structural connectivity to impinge on brain function remain challenging in part because white matter tractography algorithms for diffusion tensor imaging (DTI) may lose fidelity in the presence of lesions. We used a “virtual lesion” approach to characterize the “disconnectome” associated with periventricular white matter (PVWM) lesions. We simulated progressive ischemic PVWM lesions using sub-threshold cerebral blood flow (CBF) masks derived from a previously published group-averaged map acquired from N=436 middle aged subjects in which the lowest CBF values were seen in PVWM and morphologically recapitulated the spatial pattern of PVWM hyperintensities seen in typical aging. We mimicked the age-dependent evolution of PVWM lesion burden by varying the threshold applied to the CBF map. We found that the optic radiations, inferior fronto-occipital fasciculus, inferior longitudinal fasciculus, corpus callosum, temporopontine tract and fornix were affected in early simulated PVWM lesion burdens, and that the connectivity of subcortical, cerebellar, and visual regions were significantly disrupted with increasing simulated PVWM lesion burdens. We also validated the use of virtual lesions to simulate the disconnectome due to WM hyperintensities in a cognitively normal elderly cohort (N=46) by evaluating correlations between structural and functional connectomes. The virtual lesion approach provides new insights into the spatial-temporal changes of the brain structural connectome under progressive PVWM burdens during normal aging.Significance StatementWe determined the disconnectomes caused by periventricular white matter (PVWM) lesions using the “virtual lesion” approach. We validated the approach using lesions, DTI and resting-state fMRI data from elderly subjects. We simulated disconnectome of progressive PVWM lesions using cerebral blood flow (CBF) masks in PVWM region with normative DTI data, which provides specificity for an ischemic mechanism and begins to address the possibility that connectivity may be affected by reduced CBF prior to the development of overt lesions on T2-weighted FLAIR MRI. The current study presented new insights into the spatial-temporal evolutions of the brain structural connectome under progressive PVWM burdens under normal aging.

A Case-Control Clinical Trial on a Deep Learning-Based Classification System for Diagnosis of Amyloid-Positive Alzheimer’s Disease

“…It can also detect structural changes related to many brain diseases other than AD. Several recent studies, [8][9][10][11][12][13] including our previous work, 14 have developed artificial intelligence (AI)-based algorithms for classifying AD using structural brain MRI, with promising results in terms of processing time and classification accuracy. In the case of our previous work, our deep learning-based classification system for AD using structural brain MRI (DLCS) demonstrated excellent accuracy in classifying probable AD patients from cognitively normal controls (area under the curve [AUC]=0.88-0.94).…”