A three-step multimodal analysis framework for modeling corticomuscular activity with application to Parkinson's...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Systems Biology

Background:

• Corticomuscular coupling analysis using electroencephalography (EEG) and electromyography (EMG) is crucial for understanding human motor control.
• Conventional pair-wise magnitude-squared coherence (MSC) methods have limitations in group inference, multi-dataset analysis, and interaction assumptions.

Purpose of the Study:

• To develop an advanced method for assessing corticomuscular coupling by integrating multiset canonical correlation analysis (M-CCA) and joint independent component analysis (jICA).
• To overcome the limitations of traditional MSC methods for multi-dataset analysis and group inference.

Main Methods:

• Proposed a novel approach combining M-CCA and jICA to analyze corticomuscular coupling.
• Ensured extracted components are maximally correlated across multiple datasets and statistically independent within each dataset.
• Applied the method to concurrent EEG, EMG, and behavioral data from Parkinson's disease (PD) patients.

Main Results:

• The combined M-CCA and jICA method successfully identified highly correlated temporal patterns across EEG, EMG, and behavioral signals.
• Spatial activation patterns revealed enhanced occipital connectivity in PD subjects, aligning with existing medical findings.
• Demonstrated the method's capability to analyze complex, multi-modal neurophysiological data.

Conclusions:

• The proposed M-CCA and jICA method offers a robust and versatile tool for corticomuscular coupling analysis across multiple datasets.
• This advanced technique provides deeper insights into motor control and neurological disorders like Parkinson's disease.
• Findings highlight potential neural alterations in PD beyond motor areas, specifically in visual processing regions.