Quantifying muscle coactivation in individuals with incomplete spinal cord injury using wavelets

Abstract

BackgroundIndividuals with incomplete spinal cord injury often have decreased gait function and coactivation of antagonistic muscle pairs. Common ways of quantifying coactivation using electromyographic signals do not consider frequency information in the signal. As electromyographic signals from different motor unit types have different frequency components and muscle fiber type can change in individuals with spinal cord injury, it may be beneficial to consider frequency components. The aims were to demonstrate the utility of using a method which considers temporal and frequency components of the electromyographical signal to quantify coactivation in lower extremity muscles in individuals with incomplete spinal cord injury through 1) comparison with able-bodied individuals and 2) comparison before and after body weight supported treadmill training. MethodsFrequency decomposition techniques were applied to electromyographical signals to consider the temporal and frequency components of the electromyographical signals to quantify coactivation over a range of frequencies. ResultsOur main findings show that correlation coefficients between total EMG intensities of rectus femoris-biceps femoris and medial gastrocnemius-tibialis anterior were significantly different between able-bodied individuals and those with incomplete spinal cord injury (p = 0006, p = 0.01). The correlation spectra of medial gastrocnemius-tibialis anterior of the spinal cord injury group were substantially different than those the able-bodied group, while the EMG normalcy score was significantly different (p = 0.002). We also found that there was a change in coactivation of ankle muscles after body weight supported treadmill training. InterpretationOur findings indicate that there may be frequency specific differences in muscle coactivation between able-bodied individuals and those with incomplete spinal cord injury. Changes in coactivation were also observed before and after body weight supported treadmill training. These differences may reflect the changes in recruitment patterns of different motor unit types.