Immediate after-effects of robot-assisted gait with pelvic support or pelvic constraint on overground walking in healthy subjects

BACKGROUND: Recovery of walking is a primary rehabilitation goal of most stroke survivors. Control of pelvic movements is one of the essential determinants of gait, yet surprisingly, conventional robot-assisted gait trainers constrain pelvic movements. Novel robot-assisted gait trainers, such as LOPES II, are able to support pelvic movements during gait. The aim of this cross-over study was to investigate the immediate after-effects of pelvic support (PS) or pelvic constraint (PC) gait training with LOPES II on overground walking in healthy subjects. METHODS: Thirteen able-bodied subjects (22.8 ± 2.1 years) participated in two 20-min gait training sessions with LOPES II; one with PS and one with PC. During the PS-training, the LOPES II actively guided the lateral displacement of the pelvis, while pelvic rotations were free. During the PC-condition, both lateral displacement and pelvic rotations were constrained and reduced to a minimum. The training sessions were separated by a 30-min resting period. Lateral displacement of the pelvis, hip and knee kinematics, and spatiotemporal parameters during overground walking were determined at baseline and immediately following the training using 3D gait analysis. RESULTS: During the PS-condition in LOPES II the lateral pelvic displacement was significantly greater (105.6 ± 0 .5 mm) than during the PC-condition (10.8 ± 0 .7 mm; p < 0.001). Analysis of the first five steps of overground walking immediately following PC-condition showed significantly smaller lateral displacements of the pelvis (32.3 ± 12.0 mm) compared to PS-condition (40.1 ± 9 .8 mm; p < 0.01). During the first five steps, step width was significantly smaller after PC-condition (0.17 ± 0. 04 m) compared to PS-condition (0.20 ± 0.04 m; p = 0.01) and baseline (0.19 ± 0. 03 m; p = 0.01). Lateral displacement of the pelvis and step width post training returned to baseline levels within 10 steps. PC- nor PS-condition affected kinematics, gait velocity, cadence, stride length or stance time. CONCLUSIONS: In healthy subjects, robot-assisted gait training with pelvic constraint had immediate negative after-effects on the overground walking pattern, as compared to robot-assisted gait training with pelvic support. Gait training including support of the lateral displacement of the pelvis better resembles the natural gait pattern. It remains to be identified whether pelvic support during robot-assisted gait training is superior to pelvic constraint to promote gait recovery in individuals with neurological disorders.