Motor Unit-Driven Identification of Pathological Tremor in Electroencephalograms

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Traditional coherence analysis of electroencephalograms (EEG) and electromyograms (EMG) for tremor studies is limited by long recording times and susceptibility to artifacts.
• Pathological tremors, such as essential tremor (ET) and Parkinson's disease (PD), involve complex neural and muscular interactions.
• Existing methods struggle to accurately isolate tremor-related cortical signals, hindering a deeper understanding of tremor pathophysiology.

Purpose of the Study:

• To develop and validate a novel, artifact-independent method for extracting tremor-related electroencephalogram (EEG) components in pathological tremor.
• To improve the precision and robustness of analyzing neural mechanisms underlying tremors.
• To overcome limitations of traditional EEG-EMG coherence analysis in tremor research.

Main Methods:

• Developed a linear minimum mean square error estimator for the tremor component in EEG, leveraging the coupling with motor unit activity.
• Estimated motor unit population activity by decomposing surface electromyogram (EMG) signals into cumulative spike trains (CST).
• Initialized and optimized the tremor-related EEG component estimate using the derived CST and a novel optimization approach.

Main Results:

• The new method demonstrated robustness against noise and motor unit firing variability on simulated data across various tremor frequencies.
• In essential (ET) and Parkinson's disease (PD) patients, the method yielded approximately a two-fold increase in coherence amplitude between the estimated EEG component and CST compared to classical EEG-EMG coherence.
• The novel approach successfully tracked changes in tremor-related cortical activity over time in patient data.

Conclusions:

• A novel, robust method for precise estimation of tremor-related EEG components has been developed.
• This technique eliminates the need for artifact removal and provides reliable results from shorter EEG/EMG recordings.
• The method offers enhanced insights into the neural dynamics of pathological tremors, outperforming traditional coherence analysis.