Artifact Adaptive Ideal Filtering of EMG Signals Contaminated by Spinal Cord Transcutaneous Stimulation

Area of Science:

• Biomedical Engineering
• Neuroscience
• Signal Processing

Background:

• Electromyography (EMG) signals are crucial for understanding neuromuscular function.
• Spinal cord transcutaneous stimulation (scTS) is used in research and therapy but can introduce artifacts into EMG recordings.
• Accurate EMG analysis requires effective removal of these stimulation-induced artifacts.

Purpose of the Study:

• To characterize the nature and extent of EMG signal contamination from scTS.
• To evaluate the performance of the Artifact Adaptive Ideal Filtering (AA-IF) technique for scTS artifact removal.
• To compare AA-IF with the empirical mode decomposition Butterworth filtering (EMD-BF) method.

Main Methods:

• scTS was applied to five participants with spinal cord injury (SCI) at varying intensities and frequencies.
• EMG signals from Biceps Brachii (BB) and Triceps Brachii (TB) muscles were recorded at rest and during voluntary activation.
• Fast Fourier Transform (FFT) analyzed artifact frequencies; AA-IF and EMD-BF techniques were used for artifact removal and comparison.

Main Results:

• scTS artifacts contaminated ~2Hz wide frequency bands, influenced by stimulation intensity, muscle activation state, and muscle type.
• The width of artifact contamination increased with scTS intensity and was wider at rest and in the BB muscle compared to TB.
• AA-IF preserved significantly more of the uncontaminated EMG signal's spectral content (96±5%) than EMD-BF (75±6%).

Conclusions:

• The AA-IF technique accurately identifies and removes scTS-induced artifacts from EMG signals.
• AA-IF is superior to EMD-BF in preserving the genuine EMG signal content after artifact removal.
• This improved artifact removal enhances the reliability of EMG analysis in the presence of scTS.