Clinical Scenario

• He currently is able to ambulate for 3 minutes without needing to rest due to fatigue.
• The patient is a motivated, young, and athletic individual who is adamant about returning to walking, and hopefully skiing, as soon as possible. With such a great interest in the field of engineering, he inquires about assistive technology rehabilitation programs as a potential enhancement to his therapy plan of care.

Clinical Question:

What brain-computer interface technology is available for the purpose of improving ambulation capabilities in patients with incomplete SCI?

Search Strategy

• Inclusion criteria: patients diagnosed with SCI; use of BCI for purpose of gait-specific neurorehabilitation
• Exclusion criteria: brain-computer interface for upper extremity, non-motor/cognitive functions; BCI within only virtual reality environment; BCI using invasive procedures such as brain or spinal implants

Evidence Summary Table 

Outcomes Summary

• In Do et. al, the control or maximum value using this computerized analysis was 0.498, while the actual average cross-correlation value across the 5 sessions was 0.815. This indicates that the cross-correlations were significant with an empirical p-value of less than .001. The subject, diagnosed with a T6 ASIA B spinal cord injury, even gained accurate control of the system in their first attempt.
• A similar cross-correlation analysis was used in both King et. al studies.  The 2014 study resulted in a p-value of less than .001 for almost all experimental sessions and the 2015 study resulted in a p-value of less than .01 in all sessions. Both of these studies illustrate purposeful device control of the BCI- FES system in walking and idling states.
• Lastly, Lopez-Larraz et.al calculated the decoding accuracy of each session, by dividing the number of trails in which the BMI correctly decoded the intention of motion by the number of trials in which the subject triggered the system on their own. The average decoding accuracy for the successful sessions completed by 3 SCI subjects was about 77% across 120 trials in 4 sessions.

• The BCI-RoGO demonstrated a 100% response rate with no omissions. The false alarm rate was 2 or less per session, with an average of .8. Although few in number, false alarms do carry risk of bodily harm to the patient.
• The BCI-FES system also resulted in no omissions, but the false alarm rate was a bit higher than in the RoGO study. Fortunately, the false alarm rate decreased as the patient became more accustomed to operating the BCI system.
• In the BCI-exoskeleton study, instead of measuring the omissions and false alarms, authors reported on the number of movement triggers generated by the system during “rest” or “preparation” intervals because in this protocol, the system physically impeded the exoskeleton from moving during these time frames. In the successful sessions, movements were triggered falsely in 40% of the trials during the “preparation” interval and in 63% of the trials during the “rest” interval.

• The interventions included 2 BCI gait systems, virtual reality and traditional body-weight supported gait training techniques.
• Only 1 of the 8 patients in the study had an incomplete spinal cord injury, but all were chronic cases of 3 years or greater since time of injury.
  • All patients had previously been enrolled in traditional physical rehabilitation programs but had not exhibited any level of sensory or motor improvement in the years they were followed prior to enrollment in the study.
• In addition to manual testing, surface EMG recordings were taken and the post-study results revealed that all patients exhibited some degree of motor improvement, indicated by their ability to voluntarily control at least 1 muscle below the neurological level of injury.
  • Proximal muscles surrounding the hip joint were reported to have the most gains, however there were also improvements in the knee and ankle joints.
  • By the end of the 12 months, the  ASIA B classified patient moved to an ASIA C classification and interestingly, the 7 complete spinal cord injury patients also changed to an incomplete classification by the end of the study.
• All patients showed significant improvement in assisted walking skills over the last 5 months of training, assessed with the Walking Index for SCI.
  • The patient with an incomplete injury improved their score by 4 levels and it has been suggested by previous analyses that just 1 level of improvement on this measure may be considered a real difference in a clinical context (Burns et al, 2011).

Clinical Bottom Line:

• Recent evidence has demonstrated the feasibility of using brian-computer interface-based technology as a neurorehabilitation technique for the improvement of gait in patients with chronic spinal cord injury.
• The quality and amount of evidence available is limited.
• The evidence’s value is therefore “proof-of-concept” and the results engender areas of interest for future study, rather than provide significant findings for current rehabilitative use.
• Theory of neuroplastic effects: The authors of this research argue that these systems may bring about neuroplastic effects on residual or partially spared motor pathways in incomplete SCI patients. The hypothesis is that such technologies couple the behavioral activation of the supraspinal gait areas and the spinal cord gait central pattern generator to promote Hebbian learning, which could improve outcomes beyond those of standard gait therapy.

Application to case scenario:

• Because the patient has expressed interest in novel therapeutic gait training methods that may be more beneficial to him, the physical therapist has completed a search on the NIH website for ongoing studies for which he may be appropriate.  Due to the lack of current evidence on BCI technologies, and considering his more recent onset of injury and therefore greater potential for motor recovery, the physical therapist advocates that the patient continues with his current inpatient physical therapy in combination with his research participation.

References

1. Do AH, Wang PT, King CE, Chun SN, Nenadic Z. Brain-computer interface controlled robotic gait orthosis. J Neuroeng Rehabil. 2013;10:111-111.
2. Donati ARC, Shokur S, Morya E, et al. Long-term training with a brain-machine interface-based gait protocol induces partial neurological recovery in paraplegic patients. Sci Rep. 2016;6:30383-30383.
3. López-Larraz E, Trincado-Alonso F, Rajasekaran V, et al. Control of an ambulatory exoskeleton with a brain-machine interface for spinal cord injury gait rehabilitation. Frontiers in Neuroscience. 2016:1-15.
4. King CE, Wang PT, McCrimmon CM, Chou CCY, Do AH, Nenadic Z. Brain-computer interface driven functional electrical stimulation system for overground walking in spinal cord injury participant.Conf Proc IEEE Eng Med Biol Soc. 2014;2014:1238-1242.
5. King CE, Wang PT, McCrimmon CM, Chou CCY, Do AH, Nenadic Z. The feasibility of a brain-computer interface functional electrical stimulation system for the restoration of overground walking after paraplegia. J Neuroeng Rehabil. 2015;12:80-80.