Towards standardised use of equine surface EMG: effect of normalisation and filtering on outcome measures. Comparative Exercise Physiology

Abstract

Low-frequency noise attenuation and normalisation are fundamental signal processing (SP) methods for surface electromyography (sEMG), but are absent, or not consistently applied, to equine sEMG and their influences on outcome measures are unknown. During canter, leading hindlimb (LdH) experiences greater vertical loading than trailing hindlimb (TrH), thus differences in muscle activity are hypothesised. The purpose of this study was to determine the effect of normalisation and band-pass filtering on sEMG outcome measures, calculated from LdH and TrH data. sEMG (2,088 Hz) and 3D-kinematic (232 Hz) data were synchronously collected from right Biceps Femoris in 10 horses (9.7±2.6 years, 161.9±6.3 cm) during both canter leads (4.6±0.4 m/s). Two SP methods were applied to raw sEMG data; method 1 (M1) represents the most common SP method from equine literature (DC-offset removal, full-wave rectification). Method 2 (M2) includes additional high-pass filtering (Butterworth 4th order, 40 Hz cut-off), for artefact attenuation, and normalisation relative to maximum sEMG across all strides, to reduce inter-subject variability. Integrated EMG (iEMG), mean amplitude (MA) and peak amplitude (PA) were calculated using processed sEMG from both methods and stride duration as temporal domain. Data from LdH and TrH were grouped and compared within each method using repeated measures ANOVA. For M2, LdH was significantly greater than TrH for iEMG, MA and PA (P<0.01), whereas M1 showed no differences (P>0.05). Standardised equine SP is therefore recommended to detect differences that would otherwise have been missed. Further research is required to determine whether differences in muscle activity are due to differences in limb loading