Movement variability reflects the adaptation of the neuromuscular control system to internal or external perturbations, but its relationship to stroke-induced injury is still unclear. In this study, the multifractal detrended fluctuation analysis was used to explore the stroke-induced changes in movement variability by analyzing the joint angles in a treadmill-walking task. Eight healthy subjects and ten patients after stroke participated in the experiment, performing a treadmill-walking task at a comfortable speed. The kinematics data of the lower limbs were collected by the motion-capture system, and two indicators, the degree of multifractality (α) and degree of correlation [h(2)], were used to investigate the mechanisms underlying neuromuscular control. The results showed that the knee and ankle joint angles were multifractal and persistent at various scales, and there was a significant difference in the degree of multifractality and the degree of correlation at the knee and ankle joint angles among the three groups, with the values being ranked in the following order: healthy subjects < non-paretic limb < paretic limb. These observations highlighted increased movement variability and multifractal strength in patients after stroke due to neuromotor defects. This study provided evidence that multifractal detrended analysis of the angles of the knee and ankle joints is useful to investigate the changes in movement variability and multifractal after stroke. Further research is needed to verify and promote the clinical applications.

Gait stability is the ability to deal with small perturbations that naturally occur during walking. Changes in motor control caused by pain could affect this ability. This study investigated whether nociceptive stimulation (hypertonic saline injection) in a low back (LBP) or calf (CalfP) muscle affects gait stability. Sixteen participants walked on a treadmill at 0.94ms(-1) and 1.67ms(-1), while thorax kinematics were recorded using 3D-motion capture. From 110 strides, stability (local divergence exponent, LDE), stride-to-stride variability and root mean squares (RMS) of thorax linear velocities were calculated along the three movement axes. At 0.94ms(-1), independent of movement axes, gait stability was lower (higher LDE) and stride-to-stride variability was higher, during LBP and CalfP than no pain. This was more pronounced during CalfP, likely explained by the biomechanical function of calf muscles in gait, as supported by greater mediolateral RMS and stance time asymmetry than in LBP and no pain. At 1.67ms(-1), independent of movement axes, gait stability was greater and stride-to-stride variability was smaller with LBP than no pain and CalfP, whereas CalfP was not different from no pain. Opposite effects of LBP on gait stability between speeds suggests a more protective strategy at the faster speed. Although mediolateral RMS was greater and participants had more asymmetric stance times with CalfP than LBP and no pain, limited effect of CalfP at the faster speed could relate to greater kinematic constraints and smaller effects of calf muscle activity on propulsion at this speed. In conclusion, pain effects on gait stability depend on pain location and walking speed.