Reliability and validity of DTI-based indirect disconnection measures.

Area of Science:

• Neuroscience
• Neuroimaging
• Computational Neuroscience

Background:

• White matter connections are crucial for brain network function.
• Focal brain lesions can lead to structural disconnections, impairing cognitive abilities.
• Existing methods for quantifying disconnection are indirect and require further validation.

Purpose of the Study:

• To implement and validate a tool for quantifying structural disconnection using indirect methods.
• To assess the reliability and accuracy of this tool by comparing its predictions to direct patient diffusion tensor imaging (DTI) data.
• To evaluate the clinical utility of the method in mapping visual field defects.

Main Methods:

• Development of a tool utilizing tractography data from healthy controls to predict fiber loss.
• Inclusion of uncertainty metrics for overall and voxel-level predictions.
• Validation against direct DTI measures in 95 first-ever stroke patients.
• Application to a subset of patients with hemianopia to assess clinical utility.

Main Results:

• The tool accurately predicts fiber loss for lesions larger than a minimal size, showing high reliability.
• Predictions correlate well with direct DTI estimates from patients.
• Tract-based measures effectively mapped visual field defects, outperforming lesion localization.
• The method identified networks consistent with the visual system's known anatomy.

Conclusions:

• Indirect measures of structural disconnection are reliable and valid substitutes for direct fiber loss estimations.
• This validated method offers a robust approach to understanding lesion-induced network disruptions.
• Indirect measures, when based on high-quality control data, may be preferable to low-quality single-subject DTI.