Identifiable Patterns of Trait, State, and Experience in Chronic Stroke Recovery.

Area of Science:

• Neuroscience
• Cognitive Science
• Medical Imaging

Background:

• The healthy human brain's functional connectome exhibits remarkable stability, akin to a unique fingerprint.
• Investigating functional connectivity stability in individuals with chronic stroke is essential for understanding brain plasticity and recovery.

Purpose of the Study:

• To assess the stability of functional connectivity across different tasks and over time in chronic stroke survivors.
• To determine if machine learning can identify individuals, tasks, and therapy-induced changes based on functional connectivity patterns.

Main Methods:

• Utilized functional magnetic resonance imaging (fMRI) on 12 chronic stroke survivors over 18 weeks, including intensive aphasia therapy.
• Applied supervised machine learning (support vector machine) to classify fMRI data based on participant, task, and pre/post-therapy time points.
• Employed permutation testing to establish statistical significance for classification accuracy.

Main Results:

• Functional connectivity matrices allowed for accurate classification of participants (87.1%), tasks (68.1%), and therapy time points (72.1%).
• Classification accuracy was maintained using only the contralesional hemisphere, suggesting robust individual-specific patterns.
• Resting-state data also predicted task-based data by subject, highlighting stable individual connectivity profiles.

Conclusions:

• Individual variability in functional connectivity is a critical factor in interpreting stroke recovery mechanisms.
• Findings underscore the importance of personalized approaches in stroke rehabilitation and neuroimaging research.
• The study demonstrates the potential of machine learning in characterizing brain function stability and change in neurological disorders.