Electromyographic assessment of trunk muscle activation amplitudes during a simulated lifting task using pattern recognition techniques

This study sought to determine the patterns of neuromuscular response from 24-trunk muscle sites during a symmetrical lift and replace task. Surface electromyograms (EMG) and kinematic variables were recorded from 29 healthy subjects. Pattern recognition techniques were used to examine how activation amplitude patterns changed with the different physical demands of the task (reach, phase of movement). The results indicated that there was very little trunk and pelvis motion during the task. Three principal patterns accounted for 95% of the total variation suggesting that the measured data had a simple underlying structure of variance. ANOVA results revealed significant differences in principal pattern scores. These differences captured subtle changes in muscle recruitment strategies that most likely reflect different stability and biomechanical demands. More balanced activations (bracing) between the abdominal and back sites were observed during the lighter demands, whereas differential recruitment among the back extensor sites was more predominant in the more demanding conditions. A pattern recognition technique offers a novel method to examine the relationships among a large number of muscles and test how different work characteristics change the relationships among the muscle sites.