Myoelectric Signals Initiate Feedforward Commands for Robotic Lower Limb Prostheses

Surface electromyography electrodes can record artifact-free muscle activation patterns from residual limb muscles within the prosthetic socket-limb interface. Therefore, myoelectric signals recorded from surface electrodes within the socket-limb interface could be used in robotic lower limb prostheses to derive feedforward commands from the amputee’s nervous system, according to a recent study.

“We wanted to see what the residual muscle activation signals of lower limb amputations looked like during walking to determine if the signals could be used to control a robotic ankle,” Stephanie Huang, a PhD candidate at the University of Michigan, told O&P Business News. “Important questions had not been addressed in previous studies, including what is the muscle activation signal quality of residual muscles during walking? How consistent are the muscle activation patterns across strides? Are lower limb amputees able to learn a new residual muscle activation pattern during walking?”

EMG results

Researchers recorded surface electromyography (EMG) from seven leg muscles in the residual limb of 12 unilateral transtibial amputee patients and the right leg of 12 non-amputee patients during treadmill walking at 0.7, 1.0, 1.3 and 1.6 m/s. Lower leg muscle signals were recorded from within the limb-socket interface and from muscles above the knee for amputee patients. Patients were free of musculoskeletal and cardiovascular conditions that would limit their ability to walk safely on a treadmill, and all amputee patients had been using their prostheses for at least 6 months.

Using cross-correlation analyses, researchers quantified differences in the muscle activation profile between amputee and control groups during treadmill walking, and calculated variance-to-signal ratios to assess the step-to-step inter-subject variability profiles.

Muscle recruitment signals from residual lower leg muscles recorded within the prosthetic socket, which were locked to particular phases of the gait cycle, were reliable in amputee patients during walking. However, amputee patients had higher muscle activation profile variability vs. control patients. Upper leg muscle activation patterns were more similar between the two groups during walking, according to study results. Compared with the control group, the amputee group had significantly greater variance-to-signal ratios of EMG during 1.0 m/s treadmill walking, but, according to a post-hoc t-test, only the gastrocnemius medial head was significantly different between the groups.

“The overall results showed that in residual limbs, muscle activation signals of transtibial amputees seem like a very viable control source for mileage control during walking. Right now there is a lot of research being done on myoelectric control of lower limb prostheses. A recent study has shown that lower limb amputees can learn to control ankle position by activating their residual muscles while seated. But can amputees learn to do the same while standing or during walking?” Huang said. “To our knowledge, no one has looked at whether amputees can learn to control a robotic prosthesis using only their residual muscles during walking. When a patient is walking it’s a different story because, not only is their residual limb supporting body weight, they are performing a dynamic task.”

Future studies

The next step for Huang and colleagues is looking at signal adaptability and seeing if amputees can learn a different residual activation pattern to control prosthetic dynamics during locomotion.

“Currently, we are conducting studies to see whether lower limb amputees can learn to walk using a robotic prosthesis that is powered by pneumatic artificial muscles. The artificial muscles contract when the patient activates their residual muscles. In this way, the patient has direct control of how their ankle behaves, and for the first time in years, the amputees can see and feel their ankle moving in a way that they did before amputation,” Huang said. “Our patients are really excited to wear the experimental powered prosthesis because they realize that once they get the hang of it, the ankle does what they want it to do.”

“Thinking of clinical applications of this research is very promising,” she concluded. “Incorporating feedforward myoelectric control with current lower limb prosthetic control systems could allow amputees to perform everyday tasks that are impossible or cumbersome to perform using currently available commercial prostheses. For example, standing on your toes to reach something on a high shelf or walking down a grassy hill, things that most of us take for granted.”

For more information:
Huang S, Ferris DP. Muscle activation patterns during walking from transtibial amputees recorded within the residual limb-prosthetic interface. J Neuroeng Rehabil. Aug. 10, 2012. [Epub ahead of print]

Disclosure: Huang has no relevant financial disclosures.

Leave a Reply

Your email address will not be published.