Resistive Robotic Gait Training to Restore Neuromuscular Function in Cerebral Palsy
AuthorConner, Benjamin Charles
AdvisorLerner, Zachary F.
Kruer, Michael C.
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractIndividuals with cerebral palsy (CP) have deficits in strength and neuromuscular coordination, which contribute to a slow and inefficient gait pattern that makes walking difficult for an overwhelming majority of this population. Previous interventions seeking to address these aspects of gait dysfunction in CP have been unsuccessful to date in addressing both muscle weakness and motor control. In an effort to address this gap in clinical care, we developed and validated an adaptive control scheme in conjunction with an untethered ankle exoskeleton device to provide ankle plantar flexion resistance during walking that was responsive to user input. We found that compared to baseline, walking with adaptive plantar flexion resistance resulted in a 45 ± 35% increase in stance-phase plantar flexor activity (p = 0.02) and a 46 ± 25% reduction in stance-phase co-contraction at the ankle (p = 0.02) for children and young adults with CP. This modality was then applied in a pilot clinical trial in CP, whereby six pediatric participants completed ten, 20-minute training sessions with adaptive plantar flexion resistance over four weeks. We observed significant improvements in measures of strength (17 ± 8% increase in ankle plantar flexion strength, p = 0.02), preferred walking speed on a treadmill (39 ± 25% increase, p = 0.04), energetic efficiency (33 ± 9% reduction in metabolic cost of transport, p = 0.03), and measures of mobility (11 ± 9% improvement in timed up and go performance, p = 0.04; 13 ± 9% increase in six minute walk test distance, p = 0.04). These improvements in gross measures of performance were likely a result of observed improvements in neuromuscular control and mechanical efficiency, with training resulting in a 29 ± 11% decrease in co-contraction at the ankle (p = 0.02), a 33 ± 13% more typical soleus muscle activation profile (p = 0.01), a 7 ± 3% increase in neural control complexity (p < 0.01; measured via muscle synergy analysis), and a 58 ± 34% more mechanically efficient gait pattern (p < 0.05). Overall, this novel resistive robotic gait training paradigm demonstrated significant promise in improving strength and neuromuscular control at the ankle for improved mobility in children with CP. To further enhance the efficacy of this intervention, we developed an electrodeless audiovisual biofeedback system that utilized force sensitive resistors to display real-time plantar pressure to a user while walking. In validating this system against a soleus electromyography (EMG) biofeedback system in eight individuals with CP, which was considered the gold standard, we found comparable increases in mean soleus muscle activation relative to baseline (43 – 58%, p < 0.05), as well as mean (68 – 70%, p < 0.05) and peak (71 – 82%, p < 0.05) medial gastrocnemius activation, with strong relationships between the two systems for these outcome variables (R = 0.89 – 0.94). When this system was applied to our adaptive plantar flexion resistance scheme, it rapidly increased mean (36%, p < 0.05) and peak (46%, p < 0.05) soleus activation relative to resistance alone. The integration of this plantar pressure biofeedback system may help to improve active engagement with our resistive ankle exoskeleton scheme, reducing the necessity of constant verbal coaching or long acclimation periods. Finally, we aimed to better understand the underlying neuromuscular response to walking with our resistive ankle exoskeleton, as well as answer fundamental questions about reflex modulation in CP. We tested the effect of changes in motor task complexity, requiring varying levels of ankle stability, on soleus H-reflex excitability in this population. We found that individuals with CP displayed the typical decrease in soleus H-reflex excitability with increased standing task complexity (-26 ± 27%, p = 0.04). We also observed significant inverse relationships between soleus H-reflex amplitude and co-contraction at the ankle during both complex standing (R = -0.58, p < 0.01) and walking (R = -0.52, p < 0.01) tasks, suggesting the presence of reciprocal inhibition, which was previously thought to be absent in CP.
Degree ProgramGraduate College
Clinical Translational Sciences