Neuroprosthetics and the Making of a Cyborg Olympic Champion

Home > Articles > Neuroprosthetics and the Making of a Cyborg Olympic Champion
By Phil Stevens, MEd, CPO, FAAOP

FES Team Meltin.

Source/image rights: ETH Zurich /  Nicola Pitaro.

Last fall found me in Atlanta at the American Congress of Rehabilitation Medicine presenting on a newly completed Clinical Practice Guideline (CPG) for the orthotic management of paraplegia secondary to spinal cord injury.1 The CPG summarized the highest levels of evidence relative to this episode of care and included several succinct evidence-based recommendations. These included guidelines on treatment principles with traditional orthoses, guidance on the preliminary benefits of powered exoskeletal orthoses, the current deficiencies of hybrid solutions combining mechanical external supports with functional electrical stimulation (FES), and the inability of any of these systems to facilitate walking speeds that approach those observed in normal human locomotion.1 The document is a reasonable one, highlighting overviews of the established benefits of both traditional and emerging treatment pathways for this population.


However, the CPGs were confined to evidence statements that could be extracted from published systematic reviews that examined lower-limb orthotic literature. It summarized where things are, without any attempt to predict where things might go. This article introduces a body of research originating at the Advanced Platform Technology Center of the Louis Stokes Cleveland Veterans Affairs Medical Center (CVAMC) and Case Western Reserve University (CWRU), both in Cleveland, where comprehensive multichannel implanted neural stimulation is being used to enable standing and transfers with minimal reliance on external orthotic support, with applications well beyond these routine daily tasks. It is a compelling approach with some encouraging results that warrants awareness among clinicians that contribute to the current rehabilitation pathways for this treatment population.


The System

A very comprehensive overview of the system can be found in a 2012 publication from CWRU, CVAMC, and their collaborating partners.2 It begins by restating a basic premise well known by any clinician who has successfully fitted a patient with a WalkAide or Bioness device, namely that when neurologic damage is confined to the central nervous system, as is generally the case in spinal cord injury, the peripheral nerves are often intact and can be stimulated using electric current to create muscular contraction. Clinical orthotists are familiar with single-channel neural stimulation systems where a single surface electrode is used to generate the functional contraction of a single muscle group (e.g., the dorsiflexors of the ankle). However, these same principles can be used to generate more coordinated series of contractions across multiple muscle groups to produce more complex movement patterns such as standing and transfers, a premise that has been explored in spinal cord injury rehabilitation since the 1970s.


Surface neural stimulation is viable in single-channel systems targeting the peroneal nerve because of its relatively superficial location. However, such systems are limited by inherent challenges. Cutaneous electrical signals need to be strong enough to penetrate the poorly conductive surrounding tissues without causing damage to the skin. Further, they can only be applied to reasonably superficial motor nerves. Additionally, surface systems are cyclically donned and removed, leading to inconsistencies in the signal qualities received by the nerve. However, these collective limitations are bypassed when electrical currents are provided directly to the muscle through surgically implanted electrodes.


In the CWRU/CVAMC implanted lower-limb neural prosthesis system, eight such electrodes are utilized. Two intramuscular electrodes are implanted at the L1-L2 spinal roots to stimulate the erector spinae and provide extension of the trunk. Moving distally, two pairs of epimysial electrodes are implanted bilaterally on the gluteus maximus and semimembranosus to elicit hip extension. The final pair of electrodes are sutured bilaterally to the vastus lateralis to provide knee extension. All eight electrodes are routed to a surgically implanted pulse generator sutured to the abdominal wall. This generator ultimately provides the electrical pulses to each of the implanted electrodes using independently adjusted amplitudes, frequencies, and durations.2


Following implantation and a brief recovery period, the authors describe the creation of a profile for each electrode, determined by its threshold (the minimum pulse width that generates a visible muscle contraction) and its saturation (the pulse width above which muscle strength no longer increases or undesirable movements or sensations occur).2


With these profiles established, the patients, all of whom are several months to several years post-injury, undergo an eight-week period of muscular reconditioning, building both muscle strength and endurance. Strengthening protocols are similar to what one might encounter in other weight-lifting programs, with three sets of ten repetitions of knee extension. Contractions are held for 11 seconds, followed by 16 seconds of rest with five-minute rest periods between each consecutive set. Resistance is provided in the form of ankle weights that are progressively increased until they reach the maximal load at which all 30 repetitions can be completed.2


Endurance training is less familiar, with the patient lying supine while all electrodes are stimulated to contract, simulating the muscle patterns required for standing. These contractions are progressively increased until the subject experiences a cyclical pattern of 26 seconds of sustained contractions followed by ten seconds of relaxation for 120 minutes.2


Following reconditioning, functional therapy begins with balance and transfer training. Standing begins in a standing frame to ensure adequate muscle strength to sustain the posture, build additional endurance, and avoid the orthostatic hypotension that can occur with a return to a vertical alignment. Standing progresses through use of static parallel bars until the patient reaches the desired capacity to don lower-limb orthoses that will stabilize the feet and ankles, and to stand with the assistance of a walker.2


The stimulation patterns required to produce individual leg exercises and more complex maneuvers such as standing are ultimately programmed into an external control unit that relays both power and coordinated commands to the internal pulse generator. Users can select the desired movement patterns from a library of preprogram-med simulation patterns. With their AFOs donned, patients scroll to initiate the standing program, hold onto a walker and pre-position their bodies for the sit-to-stand transfer. After a three-second delay, an audible cue signals the initiation of increasing simulation and the users engage their upper extremities to push up with their arms and assist their bodies into a standing position.2


Early Results

The balance/transfer training described above takes time, reported at 12 weeks with the caveat that it is ultimately tailored to each individual subject. Once proficiency with the system is attained and demonstrated, patients are discharged to home use with a structured home exercise period. The obvious question that follows is simply, "What happens next?" To what extent do these skills and resources translate to utilization in a home environment?


The 2012 article describes a number of preliminary longitudinal clinical observations that begin to answer these questions. It reports on 15 individuals, one of whom is often excluded from analyses as an outlier due to a protracted time to follow-up. The subjects ranged from one year to nearly 17 years post-injury, with an average of six years between injury and implantation. Baseline assessments performed at discharge from the program were compared against follow-up assessments performed at an average of 13 months later. Given the length of the rehabilitation following initial implantation, the subjects were, on average, just over two years removed from their implantation procedures at the time of follow-up.


Knee strength, assessed through dynamometry over 12 repetitions, decreased modestly but not significantly between program discharge and follow-up. Similarly, knee endurance, assessed over 40 minutes of cyclical knee extension also declined non-significantly.2


Overall utilization was sustained, though the patterns of utilization changed over time. During the 28- day period immediately following discharge, subjects used the system an average of just under 13 days, logging an average of 12 hours of total utilization time. During the 28 days immediately preceding follow-up, this usage rate was almost identical, with subjects logging an average of just under 11½ hours of usage over just under 13 days of use. However, the time spent standing increased from an average of 40 minutes to an average of two hours 15 minutes, while the time spend exercising decreased from an average of 11 hours 22 minutes to about nine hours. The authors suggested this transition may reflect individuals incorporating the system into those activities of daily living where standing is required, with shorter but more frequent bouts of standing.2



The authors also reported on such variables as maximum standing time and body weight distribution. For the purpose of this article, it is sufficient to report that there were non-significant declines in both domains, meaning average standing times decreased and the upper extremities took on a larger percentage of weight distribution over time. However, many of the subjects retained the ability to stand for long enough to perform a standing pivot transfer, described by the authors as "the primary task for which this system was originally designed," and at follow-up body weight was supported sufficiently by the lower limbs to allow all subjects that ability to release one hand from their walker, reach above shoulder height, and manipulate objects in their environment."


The final domains reported on included the stability and survivability of the electrical components, or the consistency in motor performance with identical electrical stimulation over time and sustained functionality over time. Both parameters exceeded 90 percent, suggesting a robust performance.2


Longer-term Performance

We then fast-forward five years where a 2017 publication reports on the long-term observations of the same CWRU/CVAMC system.3 This publication reports upon an expanded cohort of 22 individuals, an average of 14½ years post-injury who were an average of 6.2 years removed from the implantation procedures and rehabilitation.


Several of the domains described earlier were repeated. For example, knee extension strength was found to be maintained between discharge and follow-up. Additionally, the parameters of maximum standing time and body weight distribution were also reported again, reinforcing trends of modest declines following discharge from rehabilitation, but maintenance of sufficient performance to allow for functional tasks (i.e., transfers and overhead reaching).


New to this publication were objective reports of functional task performance, with an 87.5 percent retention of the ability to perform the predefined tasks of standing, standing to retrieve an object from an overhead shelf, and performing a standing pivot transfer.3


Also new to this publication was the introduction of the Quebec User Evaluation of Satisfaction with assistive Technology (QUEST), a self-report instrument in which respondents rated their satisfaction on a scale from one (not satisfied at all) to five (very satisfied). The 17 subjects who completed the QUEST generally reported themselves as being "quite satisfied" with the device as indicated by a mean QUEST score of 3.88. They also reported themselves as being "very satisfied" with the services received, as indicated by an average QUEST rating of 4.54 for this consideration.


In addition to the QUEST, participants completed a custom questionnaire designed to assess their experiences with their neuroprosthetic system. This scale ran from -3 (very dissatisfied) to 3 (very satisfied). Using this questionnaire, 78 percent of the subjects indicated that the neuroprosthesis lived up to their expectations, 83 percent reported that they would go through the same process again for the same result, and 94 percent said they would recommend the experience to others. Collectively, 94 percent reported being either moderately satisfied or very satisfied with the neuroprosthesis.3


So even after an average of six years with the system, subjects with the CWRU/CVAMC system have been experiencing consistent objective and subjective benefits in their home living environment.



While the academic descriptions of the CWRU/CVAMC implanted lower-limb neural prosthesis are compelling, the system garnered more mainstream attention when it was literally put to the test at the first international Cybathlon or "Cyborg Olympics," held in Switzerland in 2016. Here, the system and a "pilot," Mark Muhn, competed against 11 other FES-driven biking teams. Led by the lead author on the described papers, Ronald Triolo, PhD, the Cleveland team won the 750-meter event in a record time of 2:58 under the muscle power of the team's spinal cord injured pilot and the contractile signals of the neuroprosthesis.4


In preparation for the event, the team added a cycling program to the exercise options of the external control unit and equipped a recumbent tricycle with sensors that detect the angles of the pilot's pedaling legs and coordinate the simulation patterns so that while one leg is pushing, the other is pulling.5


Interested readers can see Muhn's winning ride at starting at about 6:40 into the video clip.



While the current limitations of the CWRU/CVAMC implanted lower-limb neural prosthesis as described in recent literature are evident, they represent a tremendous amount of progress in the field of neuroprosthetics for this population and suggest an avenue of continued study and research for this treatment population.


Phil Stevens, MEd, CPO, FAAOP, is in clinical practice with Hanger Clinic, Salt Lake City. He can be contacted at



  1. Stevens P., C. Bertram, J. Campbell. 2017. Lower Limb Orthotic Management of Paraplegia due to Spinal Cord Injury: A Clinical Practice Guideline. Archives of Physical Medicine and Rehabilitation 10:e134.

  2. Triolo, R. J., et al. 2012. Longitudinal performance of a surgically implanted neuroprosthesis for lower-extremity exercise, standing and transfers after spinal cord injury. Archives of Physical Medicine and Rehabilitation 93:896-904.

  3. Triolo, R. J., et al. 2017. Long-term performance and user satisfaction with implanted neuroprostheses for upright mobility after paraplegia: 2- to 14- year follow-up. Archives of Physical Medicine and Rehabilitation.

  4. accessed on 12/15/2017

  5. accessed on 12/15/2017