Reshaping of Bilateral Gait Coordination in Hemiparetic Stroke Patients After Early Robotic Intervention

ORIGINAL RESEARCH published: 09 October 2018 doi: 10.3389/fnins.2018.00719 Reshaping of Bilateral Gait Coordination in Hemiparetic Stroke Patients After Early Robotic Intervention Sandra Puentes 1,2* , Hideki Kadone 2 , Hiroki Watanabe 3 , Tomoyuki Ueno 4 , Masashi Yamazaki 5 , Yoshiyuki Sankai 3 , Aiki Marushima 6 and Kenji Suzuki 3 1 Faculty of Engineering, Information and Systems, University of Tsukuba, Tsukuba, Japan, 2 Center for Innovative Medicine and Engineering, University of Tsukuba Hospital, Tsukuba, Japan, 3 Center for Cybernics Research, University of Tsukuba, Tsukuba, Japan, 4 Department of Rehabilitation Medicine, University of Tsukuba Hospital, Tsukuba, Japan, 5 Department of Orthopaedic Surgery, University of Tsukuba Hospital, Tsukuba, Japan, 6 Department of Neurosurgery, University of Tsukuba Hospital, Tsukuba, Japan Hemiparetic gait is a common condition after stroke which alters importantly the quality of life of stroke survivors. In recent years, several robotic interventions have been developed to support and enhance rehabilitation strategies for such population. The Hybrid Assistive Limb (HAL) robot suit is a unique device able to collect in real time bioelectric signals from the patient to support and enhance voluntary gait. HAL has been used before in early stages of stroke showing gait improvement after the intervention. However, evaluation of the coordination of gait has not been done yet. Coordination is a key factor for an adequate gait performance; consequently, its changes may be closely related to gait recovery. In this study, we used planar covariation to evaluate coordination changes in hemiparetic stroke patients after early HAL intervention. Before starting, impaired intersegmental coordination for the paretic and non-paretic side was evident. HAL intervention was able to induce recovery of the covariation loop shape and deviation from the covariation plane improving intersegmental coordination. Also, there was a tendency of recovery for movement range evidenced by comparison of peak elevation angles of each limb segment of the patients before and after HAL intervention, and also when compared to healthy volunteers. Our results suggest that early HAL intervention contributed to the improvement of gait coordination in hemiparetic stroke patients by reinforcing central pattern generators and therefore reshaping their gait pattern. R Edited by: Yury Ivanenko, Fondazione Santa Lucia (IRCCS), Italy Reviewed by: Tetsuro Funato, The University of Electro-Communications, Japan Fan Gao, University of Kentucky, United States *Correspondence: Sandra Puentes Specialty section: This article was submitted to Neuroprosthetics, a section of the journal Frontiers in Neuroscience Received: 27 June 2018 Accepted: 19 September 2018 Published: 09 October 2018 Citation: Puentes S, Kadone H, Watanabe H, Ueno T, Yamazaki M, Sankai Y, Marushima A and Suzuki K (2018) Reshaping of Bilateral Gait Coordination in Hemiparetic Stroke Patients After Early Robotic Intervention. Front. Neurosci. 12:719. doi: 10.3389/fnins.2018.00719 Trial registration: UMIN000022410 2016/05/23. Keywords: stroke, hemiparesis, robot suit HAL, gait coordination, early intervention Abbreviations: 6MWD, 6-Minute Walk Distance; EA, elevation angles; FAC, Functional Ambulation Classification; FIM, Functional Independence Measure; FMA, Fugl-Meyer Assessment; HAL, Hybrid Assistive Limb; max peaks, maximum peaks; min peaks, minimum peaks; max-min difference, maximum peak to minimum peak difference; np-pre, non-paretic side before HAL intervention; np-post, non-paretic side after HAL intervention; par-post, paretic side after HAL intervention; par-pre, paretic side before HAL intervention; PC2SD, PCA second component standard deviation; PC3SD, PCA third component standard deviation; PCA, Principal Component Analysis; PV2, PCA second component percentage of variance; PV3, PCA third component percentage of variance. Frontiers in Neuroscience | www.frontiersin.org 1 October 2018 | Volume 12 | Article 719 Puentes et al. Robotic Gait Reshaping in Hemiparesis gait assistance robot GAR found motor and functional recovery but it was not significantly different from regular rehabilitation; however, the extensor muscle torque improved bilaterally, being significant only for the non-paretic side (Ochi et al., 2015). These studies focused their analysis on the comparison of functional scales and gait parameters, but there are no additional measurements to evaluate coordination changes after robotic intervention. It is known that coordination impairment is an underlying cause of gait deficit after stroke (Bleyenheuft et al., 2009; Chow and Stokic, 2015); hemiparesis alters the stabilization of head and thorax which contribute to the deviation of walking trajectories and poor balance during gait generation (Lamontagne et al., 2005). Also, muscle weakness from the hemiparetic side affect the initiation of movement and proper flexion and extension of the ipsilateral hip, knee, and ankle (Quervain et al., 1996). Likewise, post-stroke patients have more difficulties regulating their walking speed, step frequency, and step length which are important elements to execute stable gait when walking in complex environments (Hak et al., 2013). However, despite unilateral brain damage, abnormal patterns of movement are also found in the non-paretic side during gait generation (Quervain et al., 1996) accompanied by increment of the bilateral kinematic variability. These changes have been associated to asymmetry and lack of coordination of the leftright stepping phase (Meijer et al., 2011). Additionally, these conditions impair the ability to avoid obstacles bilaterally, becoming more prominent under time pressure (Otter et al., 2005). Bilateral coordination is an important component of gait pattern; therefore, evaluation of its changes may help to elucidate the impact of robotic interventions. Kinematic analysis of gait has been attempted before by measuring changes in joint angles; however, the pattern of the flexion-extension angles of hip and knee joints tends to generate a large variation in inter-patients and inter-trial results (Boudarham et al., 2013) and becomes dependent on the gait speed (Borghese et al., 1996). Since the gait pattern is drastically altered after stroke due to hemiparesis and march instability, perturbing patients’ balance and speed (Quervain et al., 1996; Goldie et al., 2001); it is ideal to collect data using a method able to obtain reproducible results despite these gait characteristics. Analysis of the EA of the lower limbs has shown a stereotyped pattern in healthy volunteers despite gait pattern, speed variation, or anatomic discrepancies (Borghese et al., 1996; Ivanenko et al., 2008). The EA are calculated by the relationship of lower limb thigh, shank, and foot segments to the vertical. When plotted against each other, these angles covary describing regular loops over a pl (...truncated)