TL;DR
This paper introduces a novel motorized ankle-foot orthosis (MAFO) designed for gait rehabilitation and assessment. The MAFO effectively captures short-term muscle and kinematic adaptations during walking, demonstrating its potential as a portable and versatile tool for both diagnosing gait impairments and aiding in recovery.
What Problem Does This Paper Address?
Stroke is a leading cause of physical disability, and while robotic rehabilitation systems show promise, patient adherence to therapy remains a significant challenge. Traditional rehabilitation often focuses on the "bottom-up" approach, directly addressing the affected limb. This paper explores a "top-down" approach, aiming to influence peripheral behavior by altering brain states. Specifically, it addresses the need for effective tools that can both assist in gait training and provide objective, real-time measurements of a patient's progress. Existing robotic devices often lack portability or are limited in their ability to capture detailed biomechanical and neuromuscular data during gait.
Methods
The researchers developed a unilateral powered exoskeleton, the Motorized Ankle-Foot Orthosis (MAFO), weighing 1.1 kg. This device is controlled by a hybrid controller that combines a position controller with an admittance controller. The "assistance" parameter, ranging from 0% to 100%, modulates the contribution of each controller. A higher assistance percentage means the robot performs more of the movement. The gait pattern is pre-recorded from a healthy subject and modulated based on stride time. The admittance controller adjusts stiffness based on the gait phase to manage interaction torques.
Data acquisition involved measuring ankle joint angular position, angular velocity, and torque directly from the MAFO. Surface electromyography (sEMG) was used to record the activation of the Tibialis Anterior (TA) and Gastrocnemius Medialis (GM) muscles. Kinematic signals were sampled at 1200 Hz, and EMG signals at 2048 Hz. Data was processed, filtered, and averaged to create time-normalized gait cycles for comparison across different assistance levels.
Datasets
Not discussed in the paper. The study used data collected from three healthy subjects during treadmill walking.
Key Insights and Ideas
- Portability is Key: The MAFO's relatively low weight (1.1 kg) and potential for portability make it a more practical solution for rehabilitation compared to bulkier, artificial-muscle-powered devices.
- "Top-Down" Approach Enabled: The MAFO's ability to provide controlled assistance and gather detailed feedback aligns with the "top-down" rehabilitation strategy, which aims to influence neurological recovery.
- EMG and Kinematics Synergy: The simultaneous measurement of muscle activity (EMG) and joint movement (kinematics) provides a comprehensive picture of gait dynamics, crucial for both diagnosis and tracking rehabilitation progress.
- Assistance Level Matters: The degree of robotic assistance significantly influences both muscle activation patterns and the range of motion during gait, highlighting the MAFO's adaptability to different patient needs.
Key Results
- Gastrocnemius Medialis (GM) Activation: GM activation showed slightly higher amplitude with higher assistance levels, correlating with a larger range of motion in the gait pattern.
- Tibialis Anterior (TA) Activation: TA activation remained consistent with previous studies, showing less variation across different assistance levels, particularly during mid-stance to mid-swing phases.
- Ankle Angle and Velocity: Lower assistance levels led to a reduced range of motion and diminished ankle angular velocity, as the robot imposed less constraint on the subject's natural gait. Conversely, higher assistance resulted in greater range of motion and higher velocities.
- Torque: For lower assistance levels, the torque between the subject's foot and the robot was lower, indicating less resistance from the robot as it tried to follow the subject's movements.
Limitations and Future Work
Limitations: * The study was a pilot with only three healthy subjects, limiting the generalizability of the findings. * The experiments were conducted on a treadmill, which may not fully replicate overground walking due to the inherent reflex to walk when the "ground" moves.
Future Work: * Conducting a larger study with more subjects and a longer-term experimental protocol to observe evolution in measurements due to rehabilitation exercises. * Performing experiments overground to better emulate real-world walking conditions. * Integrating electroencephalography (EEG) to study walking intention, potentially leading to a brain-neural-computer interface for assistive devices.
Summary updated