Abstract
Introduction
ECG contamination of paraspinal EMG measurements is a known issue (1,2), with several proposed methods of correction(3,4). In addition to this some question remains to how much of an effect this contamination actually has on the EMG recordings.
Methods
From a population of 455 previously recorded EMG datasets, 33 severely contaminated sets of data were selected. These 33 datasets were analysed to produce the Half-Width, RMS, RMS Slope, RMS Intercept, MF Slope, and MF Intercept variables.
The Independent Component Analysis method was used to separate the EMG data into a series of additive subcomponents which allowed the removal of ECG contamination whilst preserving underlying EMG. The subcomponents were then reintegrated to produce the original EMG signal, minus the contamination.
The resultant signal data were analysed to produce the same outcome variables so a comparison could be made.
Results
Significant differences exist in the Half-Width, RMS, RMS Intercept and MF Intercept variables between pre and post ECG removal.
Any Paraspinal EMG study planning on making use of these variables should be aware of the ECG contamination issue and the ICA removal method described here.
Conflicts of Interest
None
Source of Funding
None
