Gait Adaptation and Biofeedback for Cerebral Palsy

Quantifying Patient-specific Changes in Neuromuscular Control in Cerebral Palsy: Adaptation and Biofeedback During Gait

National Institute of Neurological Disorders and Stroke (NINDS), Gillette Children's Specialty Healthcare, Northern Arizona University

This research aims to evaluate walking function in children with cerebral palsy (CP). The researchers want to understand how children with CP adapt and learn new ways of moving. They have previously found that measuring how a person controls their muscles is important for assessing walking ability and response to interventions. In these studies, they will adjust the treadmill belt speeds and/or provide real-time feedback to evaluate how a child can alter their movement. The feedback will include a wearable exoskeleton that provides resistance to the ankle and audio and visual cues based on sensors that record muscle activity. This research will investigate three goals: first, to measure how children with CP adapt their walking; second, to see if repeated training can improve adaptation rates; and third, to determine if individual differences in adaptation relate to improvements in walking function after training. This research will help develop better treatments to enhance walking capacity and performance for children with CP.

Prior research has shown that children with cerebral palsy (CP) use simplified motor control strategies compared to nondisabled (ND) peers, and that these differences in motor control are associated with walking function. While we can quantify motor control during activities like walking, the processes by which a child with CP adapts and learns new movement patterns are poorly understood. This research will use two paradigms to evaluate adaptation and motor learning in children with CP: walking on a split-belt treadmill and responding to multimodal biofeedback. Walking on a split-belt treadmill, which has two belts set at different speeds to induce asymmetry during walking, has been commonly used to evaluate adaptation in other clinical populations. Responding to multimodal feedback can also be used to evaluate an individual's capacity to adapt their walking pattern. This research will use a real-time multimodal feedback system that targets plantarflexor activity, a key muscle group that is often impaired in CP. Sensorimotor feedback will be provided using a lightweight, body-worn robotic device that provides adaptive ankle resistance and step-by-step audiovisual feedback will be provided based on muscle activity from the plantarflexors using a visual display and audible tone. This research will quantify adaptation rate (e.g., change in soleus activity or step length symmetry) in response to these perturbations, and observe the impact of repeated practice on walking function (e.g., change in walking speed). The specific aims are to: Aim-1: Quantify adaptation rates in children with CP. We will quantify adaptation rate in response to three perturbation experiments: split-belt treadmill walking, sensorimotor feedback, and audiovisual feedback. The primary hypotheses are that children with CP will exhibit reduced adaptation rates compared to ND peers, and that adaptation rates will be associated with function (Gross Motor Function Measure, GMFM-66). Aim-2: Determine whether adaptation rates change in response to repeated feedback training. Children with CP will be randomized into three groups: sensorimotor feedback, audiovisual feedback, and sensorimotor and audiovisual feedback. Each group will undergo six weeks of training (20-min, 2x/week). The primary hypothesis is that adaptation rates will increase with repeated exposures to feedback training. Aim-3: Determine whether walking function improves after repeated practice. Gait analysis will be performed at follow-up to evaluate whether feedback training induced motor learning and improved walking function. The primary hypotheses are that training will improve muscle, joint, and whole-body performance, with the greatest improvements from combined sensorimotor and audiovisual feedback.

Participants will complete 12 sessions (20 minutes of walking on a treadmill) over a 6-8 week period while receiving audiovisual biofeedback based on the muscle activity of their ankle plantarflexors. The visual feedback will be provided on a screen with a bar showing real-time muscle activity and the audio feedback will be a sound played when they reach the target level of muscle activity.

Participants will complete 12 sessions (20 minutes of walking on a treadmill) over a 6-8 week period while receiving sensorimotor biofeedback with an ankle exoskeleton that provides resistance to ankle plantarflexion during the stance phase of gait.

Participants will complete 12 sessions (20 minutes of walking on a treadmill) over a 6-8 week period while receiving both audiovisual and sensorimotor biofeedback. Sensorimotor biofeedback will be provided with an ankle exoskeleton that provides resistance to ankle plantarflexion during the stance phase of gait. The visual feedback will be provided on a screen with a bar showing real-time muscle activity and the audio feedback will be a sound played when they reach the target level of muscle activity from the plantarflexors.

Electromyography recordings from the plantarflexor muscles are used to provide audio feedback via a sound that plays when muscle activity is above target and a visual bar that displays real-time muscle activity.

Average stance-phase magnitude of soleus muscle activity from electromyography recording measured during gait at 1-month follow-up.

The total variance account for by one muscle synergy calculated from electromyography recordings during gait.

Assessment tool designed and evaluated to measure changes in gross motor function. Parts D & E focus on standing, walking, jumping, and running function.

Inclusion Criteria: Diagnosis of bilateral cerebral palsy that impacts both legs Gross Motor Functional Classification System Level II No surgery or lower-extremity injuries in prior 12 months No botulinum toxin injections in prior 3 months No prior selective dorsal rhizotomy surgery No history of seizures or cardiac conditions that would preclude walking on a treadmill for 20 minutes No current pain that hinders walking

De-identified participant data from gait analysis and outcome measures will be provided on a public data repository.