non-invasive brain machine interface (bmi)


Generally speaking, when someone talks about a "Brain-Machine Interface" for helping someone with neuromuscular injury or disease, they mean implanting an electronic chip in the person's cortex. This chip contains the control signal which is then routed to the interface, which could be a prosthetic limb or an electrical stimulation system that activates paralyzed muscles. However, an important point to make is that the control signal does not have to come from a chip in the brain. Indeed, neural signals conveying movement intention do not reside solely within the cerebral cortex and can be directly read out from multiple locations within the human body. These signals include: EEG signals (from electrodes above the scalp), EMG signals (from muscles via surface or intramuscular electrodes), peripheral nerve signals (from cuff or intrafascicular electrodes), EcoG signals (from epidural or sub-dural electrodes), etc. Further, movement intention can be communicated through motor functions that are usually spared in SCI, such as eye movements and speech (both of which can drive interfaces without any surgical intervention). Each of these signal types can be, and has been, tapped as a potential control source in the design of brain-machine interfaces (see references below). Consider, for example, a spinal cord injured patient. Obviously functions derived from spinal motoneurons that reside above the level of the injury are spared while those that are below the level of the injury are impaired or lost, depending on the severity of the injury. But with many spinal cord injuries, that still leaves many functions viable. For example, suppose that the injury occurs at the C5-C6 level (see the black line in the figure). Clearly, the subject will be paralyzed from the waist down and will be unable to walk. Arm movements will also be limited. However, as is typical with such injuries, a significant amount of elbow and shoulder mobility may be retained since some motoneurons projecting to muscles that control these movements remain viable. Therefore, if a method can be designed to utilize these spared EMG signals for control, it would be a lot simpler that putting a chip in the motor cortex of the individual. In fact, a device that performed exactly this function for exactly this patient population was made commercially available in 1997. It is known as the Freehand System, and rather than using spikes from a chip for device operation, it uses the output of a position sensor placed on the subject's shoulder to enable palmar and lateral grasps with modest control of grasp strength. This device has been fitted in dozens of individuals with SCI at the C5-C6 level with good clinical results (see reference below). Let us now suppose that the injury occurs at C4 where even shoulder movement may well be greatly restricted. There are still some active areas that can potentially be used for control including neck movement, facial muscles, speech, eye movements, tongue movements, etc. Some of these might not be ideal control signals, but they might work well enough to provide some function, and they certainly might be a better alternative than putting a chip in a person's brain. The bottom line is that implanting a chip in the brain should be the last option for rehabilitation, after all other control sources have been tried. To this point, we note that Stephen Hawkings, who has ALS and not spinal cord injury, has never had a chip implanted in his brain. At the point when hand movements could no longer be used to control the device through which he spoke, signals from his cheek muscles were used and are still being used.

Now let's consider a lower form of paralysis where upper body movements are mostly intact, but the subject is paralyzed from the waist down. In this case, it does not really make much sense to use a chip in the brain either. How the motor cortex contributes to locomotion in primates or lower animals is poorly understood - in fact, quadrupeds whose cerebral cortex has been entirely removed, exhibit fairly robust walking behavior due in large part to the existence of autonomous central pattern generators within the spinal cord. Therefore, the most sensible and readily available signals to co-opt for triggering actuation of locomotion, a rhythmic behavior, are the periodic physical signatures of the gait cycle itself, such as joint kinematics, muscle EMG, and ground reaction forces, all of which can be obtained without the need for intense surgical intervention. These variables are currently being used in exoskeleton-guided approaches that help restore movement function in individuals with SCI through non-invasive spinal stimulation (see reference).

  1. Meng, J., Zhang, S., Bekyo, A., Olsoe, J., Baxter, B., and He, B. (2016). Noninvasive electroencephalogram based control of a robotic arm for reach and grasp tasks. Scientific Reports, 38565.
  2. Kuiken, T. A., Miller, L. A., Lipschutz, R. D., Lock, B. A., Stubblefield, K., Marasco, P. D., Zhou, P., and Dumanian, G. A. (2007). Targeted reinnervation for enhanced prosthetic arm function in a woman with a proximal amputation: a case study. Lancet, 9559, 371-380.
  3. Taylor, P., Esnouf, J., and Hobby, J. (2002). The functional impact of the Freehand System on tetraplegic hand function. Clinical Results. Spinal Cord, 11, 560-566.
  4. Burridge, J. H., Haugland, M., Larsen, B., Pickering, R. M., Svaneborg, N., Iversen, H. K., Christensen, P. B., Haase, J., Brennum, J., and Sinkjaer, T. (2007). Phase II trial to evaluate the ActiGait implanted drop-foot stimulator in established hemiplegia. Journal of Rehabilitation Medicine, 3, 212-218.
  5. Rossini, P. M., Micera, S., Benvenuto, A., Carpaneto, J., Cavallo, G., Citi, L., Cipriani, C., Denaro, L., Denaro, V., Di Pino, G., Ferreri, F., Guglielmelli, E., Hoffmann, K. P., Raspopovic, S., Rigosa, J., Rossini, L., Tombini, M., and Dario, P. (2010). Double nerve intraneural interface implant on a human amputee for robotic hand control. Clinical neurophysiology: official journal of the International Federation of Clinical Neurophysiology, 5, 777-783.
  6. Schalk, G., and Leuthardt, E. C. (2011). Brain-computer interfaces using electrocorticographic signals. IEEE reviews in biomedical engineering, 4, 140-154.
  7. Abbott, W. W., and Faisal, A. A. (2012). Ultra-low-cost 3D gaze estimation: an intuitive high information throughput compliment to direct brain-machine interfaces. Journal of neural engineering, 4, 046016.
  8. Gad, P. N., Gerasimenko, Y. P., Zdunowski, S., Sayenko, D., Haakana, P., Turner, A., Lu, D., Roy, R. R., and Edgerton, V. R. (2015). Iron 'ElectriRx' man: Overground stepping in an exoskeleton combined with noninvasive spinal cord stimulation after paralysis. Conference proceedings: Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference, 1124-1127.



General correspondence should be addressed to:

Charlotte Potak
Administrative Assistant
Bizzi Lab, 46-6189
43 Vassar Street
Cambridge MA 02139

phone: (617) 253-5687
fax: (617) 258-5342