A normative modeling approach to quantify white matter changes and predict functional outcomes in stroke patients

Area of Science:

• Neuroimaging
• Neurology
• Biostatistics

Background:

• Stroke heterogeneity poses challenges for neuroimaging studies.
• Individualized assessment is crucial for understanding stroke-induced changes.
• Normative models offer a method to measure deviations from healthy brain distributions.

Purpose of the Study:

• To evaluate stroke-induced white matter microstructural abnormalities using a normative model.
• To identify potential prognostic biomarkers for stroke recovery.
• To assess group and individual-level deviations in white matter integrity.

Main Methods:

• Diffusion-weighted imaging and clinical assessments in 46 stroke patients and 46 controls.
• Automated fiber quantification for intergroup alterations in 20 fiber tracts.
• Mahalanobis distance tractometry to quantify individual deviations (MaDDLs) in 7 tracts.

Main Results:

• Significant white matter microstructural disruptions were found in specific tracts, including the corticospinal tract.
• MaDDL metrics correlated with initial functional impairment, with MaDDL_multi showing the strongest association.
• MaDDL_multi significantly improved predictive efficacy for stroke outcomes compared to traditional metrics.

Conclusions:

• Mahalanobis distance-based deviation load (MaDDL_multi) is a valuable tool for assessing stroke-related behavioral disorders and predicting prognosis.
• This normative modeling approach has significant implications for personalized clinical decision-making in stroke recovery.
• The method shows promise for application in diverse neurological diseases.