Downsampling Effects on Needle EMG Signals: A Systematic Workflow

A new study from arXiv presents a workflow to evaluate how downsampling affects needle electromyography (nEMG) signals used for detecting neuromuscular diseases (NMDs). The research addresses computational challenges posed by high and heterogeneous sampling rates in near real-time analysis. The workflow combines shape-based distortion metrics, classification outcomes from feature-based machine learning models, and feature space analysis to quantify information loss. Using a three-class NMD classification task, the study examines how different downsampling algorithms and factors impact waveform integrity and predictive performance. The findings aim to guide optimal downsampling strategies for clinical applications.