Stroke-Related Changes in the Complexity of Muscle Activation during Obstacle Crossing Using Fuzzy Approximate Entropy Analysis

This study investigates the complexity of the electromyography (EMG) of lower limb muscles when performing obstacle-crossing tasks at different heights in post-stroke subjects versus healthy controls. Five post-stroke subjects and eight healthy controls were recruited to perform different obstacle-crossing tasks at various heights (randomly set at 10%, 20%, and 30% of the leg’s length), and EMG signals from the lower limb muscles were recorded during this challenge. The fuzzy approximate entropy (fApEn) approach was used to analyze the complexity of the EMG signals recorded from the biceps femoris (BF), rectus femoris (RF), medial gastrocnemius (MG), and tibialis anterior (TA) on both sides of the body when subjects performed different obstacle-crossing tasks at various heights. Significantly smaller fApEn values were observed in the RF of the trailing limb during the swing phase in post-stroke subjects compared with healthy controls (p<0.05), which may be caused by the smaller number and less frequent firing rates of the motor units. However, during the swing phase, there were nonsignificant increases in the fApEn values of BF and RF in the trailing limb of the stroke group compared with those of healthy controls, resulting in a coping strategy when facing challenging tasks. The fApEn values that increased with height were found in the BF of the leading limb during the stance phase and in the RF of the trailing limb during the swing phase (p<0.05). The reason for this may have been a larger muscle activation correlating with the increase in obstacle height. This study demonstrates a suitable and non-invasive method to evaluate muscle function after a stroke.