The untethered exoskeletal robot provides patients with the freest and realistic walking experience by assisting them based on their intended movement. However, few previous studies have reported the effect of robot-assisted gait training (RAGT) using wearable exoskeleton in children with cerebral palsy (CP). This pilot study evaluated the effect of overground RAGT using an untethered torque-assisted exoskeletal wearable robot for children with CP. Three children with bilateral spastic CP were recruited. The robot generates assistive torques according to gait phases automatically detected by force sensors: flexion torque during the swing phase and extension torque during the stance phase at hip and knee joints. The overground RAGT was conducted for 17~20 sessions (60 min per session) in each child. The evaluation was performed without wearing a robot before and after the training to measure (1) the motor functions using the gross motor function measure and the pediatric balance scale and (2) the gait performance using instrumented gait analysis, the 6-min walk test, and oxygen consumption measurement. All three participants showed improvement in gross motor function measure after training. Spatiotemporal parameters of gait analysis improved in participant P1 (9-year-old girl, GMFCS II) and participant P2 (13-year-old boy, GMFCS III). In addition, they walked faster and farther with lower oxygen consumption during the 6-min walk test after the training. Although participant P3 (16-year-old girl, GMFCS IV) needed the continuous help of a therapist for stepping at baseline, she was able to walk with the platform walker independently after the training. Overground RAGT using a torque-assisted exoskeletal wearable robot seems to be promising for improving gross motor function, walking speed, gait endurance, and gait efficiency in children with CP. In addition, it was safe and feasible even for children with severe motor impairment (GMFCS IV).

Pediatric gait disorders are often chronic and accompanied by various complications, which challenge rehabilitation efforts. Here, we retrospectively analyzed the feasibility of overground robot-assisted gait training (RAGT) using a joint-torque-assisting wearable exoskeletal robot. In this study, 17 children with spastic cerebral palsy, cerebellar ataxia, and chronic traumatic brain injury received RAGT sessions. The Gross Motor Function Measure (GMFM), 6-min walk test (6 MWT), and 10-m walk test (10 MWT) were performed before and after intervention. The oxygen rate difference between resting and training was performed to evaluate the intensity of training in randomly selected sessions, while the Quebec User Evaluation of Satisfaction with assistive Technology 2.0 assessment was performed to evaluate its acceptability. A total of four of five items in the GMFM, gait speed on the 10 MWT, and total distance on the 6 MWT showed statistically significant improvement (p < 0.05). The oxygen rate was significantly higher during the training versus resting state. Altogether, six out of eight domains showed satisfaction scores more than four out of five points. In conclusion, overground training using a joint-torque-assisting wearable exoskeletal robot showed improvement in gross motor and gait functions after the intervention, induced intensive gait training, and achieved high satisfaction scores in children with static brain injury.

Walking on sloped surfaces requires additional effort; how individuals with spastic hemiplegic cerebral palsy (CP) manage their gait on slopes remains unknown. Herein, we analyzed the difference in gait adaptation between the affected and unaffected legs according to changes in the incline by measuring spatiotemporal and kinematic data in children with spastic hemiplegic CP. Seventeen children underwent instrumented three-dimensional gait analysis on a dynamic pitch treadmill at an incline of +10° to −10° (intervals of 5°). While the step length of the affected legs increased during uphill gait and decreased during downhill gait, the unaffected legs showed no significance. During uphill gait, the hip, knee, and ankle joints of the affected and unaffected legs showed increased flexion, while the unaffected leg showed increased knee flexion throughout most of the stance phase compared with the affected leg. During downhill gait, hip and knee flexion increased in the affected leg, and knee flexion increased in the unaffected leg during the early swing phase. However, the ankle plantar flexion increased during the stance phase only in the unaffected leg. Although alterations in temporospatial variables and joint kinematics occurred in both legs as the slope angle changed, they showed different adaptation mechanisms.