Enhancing cognitive performance prediction by white matter hyperintensity connectivity assessment

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Vascular Neurology

Background:

• White matter hyperintensities (WMH) are linked to cognitive impairment but their volume alone doesn't fully explain deficits.
• The mechanisms connecting WMH to cognitive decline are not fully understood.
• Lesion network mapping (LNM) offers a novel approach to investigate brain network involvement in lesion-induced dysfunction.

Purpose of the Study:

• To determine if LNM-informed markers predict cognitive performance better than WMH volume.
• To identify specific brain networks associated with WMH contributing to cognitive impairment.

Main Methods:

• Utilized cross-sectional data from 3485 patients across 10 memory clinic cohorts (Meta VCI Map Consortium).
• Applied LNM to map WMH connectivity to structural and functional brain connectome data.
• Compared predictive capacity of WMH volumes and LNM scores for cognitive function using ridge regression and nested cross-validation.

Main Results:

• LNM scores significantly outperformed WMH volumes in predicting attention/executive function, processing speed, and verbal memory.
• LNM did not improve prediction for language functions.
• Higher LNM scores in attention networks correlated with poorer cognitive performance, indicating network connectivity is crucial.

Conclusions:

• WMH-related brain network connectivity measures enhance prediction of cognitive performance in memory clinic patients.
• Network integrity, especially in attention regions, is vital for understanding vascular contributions to cognitive impairment.
• Integrating connectivity data with WMH assessment may enable personalized interventions and identify at-risk patient subgroups.