Individual adjustments in patient-cooperative robotic gait training

Aims: New human-centered robotic approaches have been developed that can be applied for gait rehabilitation of patients with brain or spinal cord injury. So-called patient-cooperative controllers can take into account the patient's intention and efforts rather than imposing any predefined movement. Audiovisual displays in combination with the robotic device can be used to present a virtual environment and let the patient perform different gait and movement tasks. It is hypothesized that such human-centered robotic approaches can improve patient motivation and the quality of the therapy compared to conventional approaches.

Methods: Patient-cooperative strategies were developed and tested using the robotic gait orthosis Lokomat.

Results: Experiments with cerebral palsy and incomplete spinal cord injured patients showed that the multimodal robotic task can be adjusted to the individual patient's ability and preference. Gait speed was automatically detected either based on the horizontal propulsion force produced by the subject or based on the interaction torques between subject and robot. The treadmill training speed was adjusted accordingly. Both approaches reliably detect and control gait speed and allow for automated adjustments of training regimen to the patients ability (Fig.1 and 2, data from two healthy subjects).

Conclusion: These algorithms can be used to realize patient-cooperative training devices. Further clinical work is required to define clear, reliable and repeatable adjustment rules for the individual subjects with different CNS pathologies.