Cross-frequency Mutual Information for Cortico-muscular Coupling Detection

TY - CHAP

T1 - Cross-frequency Mutual Information for Cortico-muscular Coupling Detection

AU - Li, Xiaotong

AU - Guo, Zhenghao

AU - Wang, Cui

AU - McClelland, Verity

AU - Cvetkovic, Zoran

AU - Cong, Fengyu

PY - 2025/4/8

Y1 - 2025/4/8

N2 - Cortico-muscular coherence (CMC) is widely used to detect the interactions between the brain and muscles. How- ever, traditional CMC analysis only captures linear associations between processes and often lacks sensitivity. As a result, some healthy individuals with good motor skills exhibit no significant CMC. This study proposes a new approach, Mutual Infor- mation in Modes (MIM), which integrates Variational Mode Extraction (VME) and Mutual Information (MI) to detect both linear and nonlinear cortico-muscular communications. The method utilizes VME to adaptively extract different modes that reflect underlying neural oscillations from electroencephalogra- phy (EEG) and electromyography (EMG) signals. MI is then employed to detect and quantify dependencies both within and across these functional modes. We assess the feasibility of the MIM method by applying it to simulated data. Results using neurophysiological data demonstrate that MIM effectively captures the frequency-specific interactions between EEG and EMG signals, especially nonlinear cross-frequency couplings across different functional bands. The proposed methodology offers a promising tool for advancing our understanding of motor control and neurophysiological processes more generally

AB - Cortico-muscular coherence (CMC) is widely used to detect the interactions between the brain and muscles. How- ever, traditional CMC analysis only captures linear associations between processes and often lacks sensitivity. As a result, some healthy individuals with good motor skills exhibit no significant CMC. This study proposes a new approach, Mutual Infor- mation in Modes (MIM), which integrates Variational Mode Extraction (VME) and Mutual Information (MI) to detect both linear and nonlinear cortico-muscular communications. The method utilizes VME to adaptively extract different modes that reflect underlying neural oscillations from electroencephalogra- phy (EEG) and electromyography (EMG) signals. MI is then employed to detect and quantify dependencies both within and across these functional modes. We assess the feasibility of the MIM method by applying it to simulated data. Results using neurophysiological data demonstrate that MIM effectively captures the frequency-specific interactions between EEG and EMG signals, especially nonlinear cross-frequency couplings across different functional bands. The proposed methodology offers a promising tool for advancing our understanding of motor control and neurophysiological processes more generally

M3 - Conference paper

BT - EMBC 2025 - 47th Annual International Conference of the IEEE Engineering in Medicine and Biology Society

ER -