Intralimb Coordination and Intermuscular Coherence in Walking After Stroke

Abstract

Purpose: Following stroke, reduced motor control may lead to walking impairment with subsequent limitations in community participation and quality of life. Kinematic variability may reveal changes to motor control of the paretic limb compared to the non-paretic limb and may relate to walking performance. Frequency domain characteristics of the EMG reflects the activity of motor neuronal pools and the degree of synchronization, estimated as intermuscular coherence, between motor units of separate muscles. Together, kinematic measures of coordination may be a function of common neural drive to motor neuronal pools. Interlimb symmetry of stance time and average knee-ankle angle over stride may also reflect motor control. The purpose of this work is to characterize motor control in walking performance post-stroke. Methods: Twenty chronic stroke participants with mild to moderate walking impairment were recruited and completed a treadmill walking trial at preferred walking speed for up to 5 minutes. Kinematic data were acquired over the pelvic and lower extremity and EMG data was captured over the vastus lateralis and tibialis anterior bilaterally. The primary clinical measure was 10 meter walk time. Additional measures were Timed up and Go test, and the Stroke Impact Scale 3.0. Variability of sagittal plane knee-ankle angle was calculated over an average of 76 strides. Results: Knee-ankle angle-angle variability was greater on the paretic limb than the non-paretic limb (p=0.002) with greater variability in swing phase than in stance phase (p=0.001). Paretic swing variability relates to 10MW time (p=0.035) and lower self-reported motor function (p= 0.015) (SIS). Stance time was greater on the non-paretic limb than the paretic limb (p=0.019) and stance asymmetry related to all clinical measures. The difference between mean cyclograms of paretic and non-paretic limbs did not relate to any clinical measure. Asymmetry ratio and paretic swing variability were the greatest predictors of 10MW time. Median frequency of the tibialis anterior was lower on the paretic limb compared to the non-paretic limb (p=0.009). Within the group data, there were no differences in intermuscular coherence between the paretic and non-paretic limb and no relationship between intermuscular coherence and clinical measures. However, 13 of 17 participants showed differences in intermuscular coherence between limbs with 6 participants showing greater coherence on the paretic limb, 6 with greater coherence on the non-paretic limb and 1 with mixed results between stance and swing phase. Within the pooled data, intermuscular coherence was greater in the non-paretic limb than the paretic limb (p=0.023). Conclusion: Interlimb symmetry and knee-ankle variability relate to walking performance. However, interlimb angle-angle asymmetry does not relate to walking performance. Frequency domain characteristics between the non-paretic and paretic limb are unclear as differences are not present in the group data but are present in pooled data. The relationship between intermuscular coherence and walking performance may require more detailed characterization of bilateral stance-swing dynamics in order to meaningful relate to 10MW time.