Changes in leg cycling muscle synergies after training augmented by functional electrical stimulation in subacute stroke survivors: a pilot study

Abstract

Background: Muscle synergies analysis can provide a deep understanding of motor impairment after stroke and of changes after rehabilitation. In this study, the neuro-mechanical analysis of leg cycling was used to longitudinally investigate the motor recovery process coupled with cycling training augmented by Functional Electrical Stimulation (FES) in subacute stroke survivors.

Methods: Subjects with ischemic subacute stroke participated in a 3-week training of FES-cycling with visual biofeedback plus usual care. Participants were evaluated before and after the intervention through clinical scales, gait spatio-temporal parameters derived from an instrumented mat, and a voluntary pedaling test. Biomechanical metrics (work produced by the two legs, mechanical effectiveness and symmetry indexes) and bilateral electromyography from 9 leg muscles were acquired during the voluntary pedaling test. To extract muscles synergies, the Weighted Nonnegative Matrix Factorization algorithm was applied to the normalized EMG envelopes. Synergy complexity was measured by the number of synergies required to explain more than 90% of the total variance of the normalized EMG envelopes and variance accounted for by one synergy. Regardless the inter-subject differences in the number of extracted synergies, 4 synergies were extracted from each patient and the cosine-similarity between patients and healthy weight vectors was computed.

Results: Nine patients (median age of 75 years and median time post-stroke of 2 weeks) were recruited. Significant improvements in terms of clinical scales, gait parameters and work produced by the affected leg were obtained after training. Synergy complexity well correlated to the level of motor impairment at baseline, but it did not change after training. We found a significant improvement in the similarity of the synergy responsible of the knee flexion during the pulling phase of the pedaling cycle, which was the mostly compromised at baseline. This improvement may indicate the re-learning of a more physiological motor strategy.

Conclusions: Our findings support the use of the neuro-mechanical analysis of cycling as a method to assess motor recovery after stroke, mainly in an early phase, when gait evaluation is not yet possible. The improvement in the modular coordination of pedaling correlated with the improvement in motor functions and walking ability achieved at the end of the intervention support the role of FES-cycling in enhancing motor re-learning after stroke but need to be confirmed in a controlled study with a larger sample size.

Trial registration: ClinicalTrial.gov, NCT02439515. Registered on May 8, 2015, .

Keywords: Cycling; Electromyography; Functional electrical stimulation; Lower limb; Muscle synergy; Rehabilitation; Stroke.

Conflict of interest statement

The authors declare that they have no competing interests.

Figures

Fig. 1

Experimental setup used for FES-cycling training. The intervention lasted 3 weeks (15 sessions) and each daily session consisted of 60 min of usual care and 25 min of FES-cycling training. FES-cycling required the use of a cycle-ergometer, force sensors at the pedals, and a neuromuscular electrical stimulator connected to 8 muscles (4 for each legs). The figure shows also the visual feedback displayed to the subject during training and the FES electrodes placement (on the right)

Fig. 2

Synergy weights and activation coefficients of the four extracted synergies (Syn1, Syn2, Syn3, Syn4). Muscle synergies extracted before (T1, in light red) and after the training (T2, in dark red) for subject P8 and subject P9 are shown in panel (a) and (b), respectively. In all panels, the mean synergy weights and activation coefficients obtained by the healthy group [11] are shown in green. In the right panels, the similarity with the healthy weights obtained at T1 and T2 is reported. The left panels show the activation coefficients as function of the pedaling cycle (the grey area represents the knee flexion phase)

Fig. 3

Median [IQR] Variance Accounted For of Non-negative Matrix Reconstruction with WHEALTHY at 30 RPM. * p < 0.05

Fig. 4

Changes in spatial and timing component of muscle synergies after training. For each muscle synergy obtained by the affected leg, spatial change, i.e. one minus the similarity between the weights extracted at T1 and T2, was plotted against timing change, i.e. one minus the SSI between activation profiles at T1 and T2 after reconstruction with pre-intervention weights. Each patient is reported with a different symbol. Green and red symbols represent patients having a gait speed change after training ≥16 cm/s or