Language-related functional connectivity in post-stroke aphasia: preliminary findings from a graph-theoretical and...

Area of Science:

• Neuroscience
• Cognitive Science
• Medical Imaging

Background:

• Post-stroke aphasia (PSA) significantly impacts communication abilities.
• Understanding language network connectivity is crucial for effective rehabilitation strategies.

Purpose of the Study:

• To investigate language-related functional connectivity in PSA using graph theory during an auditory-verbal task.
• To differentiate between high-level and low-level language task performance in PSA patients.

Main Methods:

• Fifty-seven PSA patients were divided into high-level and low-level groups based on task performance.
• Functional connectivity was analyzed using ROI-to-ROI generalized psychophysiological interactions.
• Graph metrics were extracted for language nodes, and differences were assessed using ANCOVA and machine learning classification.

Main Results:

• Random Forest classification achieved significant accuracy (0.667) and AUC (0.671) in differentiating task levels.
• The left temporo-occipital inferior temporal gyrus and right posterior supramarginal gyrus were key predictors.
• High-level task performance correlated with increased local efficiency in the bilateral posterior supramarginal gyrus and decreased global efficiency in the left inferior temporal gyrus.

Conclusions:

• High-level language processing in PSA involves synergistic interaction between ventral (inferior temporal gyrus) and dorsal (supramarginal gyrus) streams.
• Increased local specialization and compensatory recruitment of the right posterior supramarginal gyrus are vital for functional recovery.
• Integrating graph metrics with interpretable machine learning shows promise for objective monitoring in aphasia rehabilitation.