Examining cortico-muscular connectivity methods on mobile brain imaging data: A comparative study with Parkinson’s Disease patients

To capture the relationship between the cortico-cortical and cortico-peripheral regions, as well as evaluate the brain functioning collectively, connectivity analysis has been progressively used. This work aims to evaluate the usability and temporal resolution of cortico-muscular connectivity methods by analyzing simultaneously recorded 128-channel electroencephalographic (EEG) and 64-channel high-density electromyographic (EMG) data from 40 participants, encompassing both healthy individuals and those with Parkinson’s disease. Participants underwent measurements during resting state (with eyes open) and while engaged in two mobile tasks – an isometric knee extension and a semi-tandem stance. EMG electrodes were placed on Vastus lateralis and Tibialis anterior, respectively. The analysis pipeline involves preprocessing steps such as filtering (0.5 Hz to 80 Hz, with additional line noise removal), manual cleaning, channel interpolation, average re-referencing, and independent component analysis decomposition. Subsequently, the EEG data is projected onto 68 cortical regions defined by the Desikan-Killiany atlas, employing the Standardized Low-Resolution Tomography (sLoreta) method. Analogously, EMG data is decomposed to extract motor unit activity using the Convolution Kernel Compensation method. Following data preparation, connectivity analysis will be conducted, with a focus on comparing standard coherence analysis to more advanced transfer entropy and hidden Markov modeling (using Gaussian observation model) methods, both on channel-level and source-projected data. Additionally, the study will explore variations in the time intervals over which the connectivity methods are calculated. The critical comparison of three distinct connectivity methods on cortico-muscular data will help in advancing further connectivity analyses.