Mind the gap: Probing the brain-machine interface
March 20, 2015
By Jeff Hansen
Experts from around the country gathered at UAB to discuss breakthroughs in linking thought with actions through technology.
On the ceiling of the Sistine Chapel, God reaches out to touch Adam. The eye locks on the small gap between their index fingers, a slice of space dividing the ethereal from man.
A similar gap challenges researchers who work with paralyzed or handicapped patients — how to synch the human brain with a machine or computer.
Surprising breakthroughs from decades of basic science were featured at the recent, all-day 2015 Bevill Neuroscience Symposium — “Brain-Machine Interfaces” at UAB, which drew 200 attendees.
The most vivid example was Jan Scheuermann, a quadriplegic from Pittsburgh. Surgeons cut through her skull in 2012 to place two 96-electrode grids on the surface of her brain. Each electrode, about the size of a pea, penetrates 1/16 inch into the brain to detect electrical signals from individual neurons. Those signals go to a decoder outside her body. The decoder translates nerve firings into commands that move a robot arm as Jan merely thinks of moving her own, paralyzed arm.
Jan Scheuermann. Photo credit: University of Pittsburgh Medical CenterUnder command of that small spot on Jan’s cortex, the machine hums into action, picking up an object in its robot hand and moving it to a shelf or table.
At first the researchers kept the arm across the room, worried that Jan might inadvertently punch herself. As Jan gained skill, the University of Pittsburgh researchers brought the arm near Jan’s side to let the paralyzed woman fulfill her greatest wish.
In meet. Neuroprosthetics like Jan’s robot arm can overcome the loss of movement. As UAB College of Arts and Sciences Dean Robert E. Palazzo, Ph.D., has explained, outside signals can also be sent into the brain. “The promise for the development of integrated portable devices that can overcome loss of hearing … and possibly even loss of sight is upon us.”
UAB’s Mark Bolding, Ph.D., mentioned even broader potentials.
The ability to feed signals back into the brain could allow someone like Jan to have proprioreception (the sense of how our bodies are positioned) or haptics (tactile sensation) from her robot arm. It could create social and affective prosthetics that create automated recognition of faces or expressions. It could be used to manage pain and modulate awareness. It could even create completely new senses — such as the ability to detect sonar, infrared light or magnetic fields, in effect allowing us to navigate like a bat, see like a bee or migrate like a bird.
“The mind-machine interface is the future of human neuroscience, using technology to overcome injury and disease of the nervous system, and at the same time providing a window into how the brain works.” — David G. Standaert, M.D., Ph.D., John N. Whitaker Professor and chair, UAB Department of Neurology
Here are overviews of the Bevill symposium talks:
- Andrew Schwartz, Ph.D., University of Pittsburgh, the speaker who showed the Jan Scheuermann video, explained the decades of science that led to her remarkable neuroprosthesis. “When you move your hand and arm, you see the whole brain light up in a complex pattern,” he said. The breakthrough came by watching the firing rate of a single neuron as a monkey moved its hand along a two-dimensional, virtual plane. “We saw that one neuron fires as the hand goes up and is silent when the hand goes down. There is an ordered relation between the rate of neuron firing and the direction of motion.”
Different neurons, it turned out, had firing rates that related to different directions. Moreover, the firing rate of each single neuron could be written as a vector dot product and those dot products could be summed into a population vector, which in neuroscience is the sum of the preferred weighted directions of a population of neurons.
The final equation to control the arm had to include three dimensions for the x, y and z axes, one aperture dimension to open and close the robot hand, three dimensions for the yaw, pitch and roll of the wrist, and four more controls that could produce 80 percent of a person’s hand shapes.
Schwartz also showed a “60 Minutes” interview with Scheuermann, where she ended the interview with a robot-arm fist bump with the CBS reporter.
In response to questions after his talk, Schwartz said there is active research on ways to activate the sensory cortex, that the neuron signals from the electrode grids tended to degrade from about 6 months to 2 years, and that figuring out how to vary force of the grip or the arm is an important topic of research.
- Leigh R. Hochberg, M.D., Ph.D., Brown University, Massachusetts General Hospital, Harvard Medical School, described the BrainGate2 Pilot Trial that also uses a small array of electrodes the size of a baby aspirin, placed on the surface of the brain.
BrainGate has 5.4 years of cumulative experience with nine people. Hochberg showed a video of one woman, 15 years after her paralyzing stroke, reaching out with a robot arm to give herself a sip of cinnamon latte.
Hochberg noted that all the recent advances and efforts — such as helping a “locked in” patient control typing or compose computer speech, controlling a prosthetic arm, reconnecting the brain with the body’s own neuromuscular system, or rehabilitating and repairing neural networks — have been built on 50 years of fundamental, publicly and philanthropically funded science. He also said that progress requires an interdisciplinary team of neurologists, neuroscientists, engineers, computer scientists, neurosurgeons, mathematicians and others.
He also mentioned potential applications of deep brain stimulation (DBS) for obsessive compulsive disorder (OCD); bipolar disorders; or to understand, predict, warn or suppress epilepsy. - P. Hunter Peckham, Ph.D., Case Western Reserve University, described implantable neural prostheses — a different approach to restorative therapy for paralyzing central nervous system disorders.
Sensors implanted under the skin near muscles that the patient can still control, such as in the shoulder or neck, pick up signals from the nerves. These go to a control device that then sends signals via other surgically implanted wires to electrodes surgically placed on the muscles of the paralyzed arms and hands. In effect, the patient’s own muscles become rewired.Peckham showed a video of a woman with a C6ASIA A spinal cord injury. Thanks to her implanted neural prosthesis, she could pick up a hair drier with two hands and blow her hair, and then brush her head with a hairbrush. Other areas in need of this kind of control are breathing (especially, to create a stronger cough), using the bladder and bowel, reaching and grasping, supporting the trunk, and standing or walking
Between 1986 and 2001, 220 spinal cord injury patients received generation 1 implants, Peckham said. Twenty patients have received the multichannel generation 2 implants that allow two simultaneous types of control, such as hand and bladder, hand and trunk, hand and standing system, or bilateral hands. Peckham said that patients will begin to get fully implantable, generation 3 neural prostheses this spring.
- Adrienne Lahti, M.D., UAB, suggested that the brain-machine interface might offer future therapies for schizophrenia.
Schizophrenia affects about 1 percent of the world’s population. While psychoses can be treated with drugs — which have side effects and require weekly blood draws — there are no treatments for cognitive and negative symptoms.
There are now multiple reports of functional connectivity abnormalities in schizophrenia, Lahti said, where the coherence between brain regions is diminished. Also, post-mortem studies of the hippocampus of schizophrenia patients show altered GABA interneurons. Furthermore, in people with OCD, deep brain stimulation of the brain’s nucleus accumbens has been shown to improve the brain’s frontostriatal functional connectivity and decrease the symptoms of OCD.
These three lines of evidence suggest a thought experiment for schizophrenia, Lahti said. Namely, can we restore connectivity, and how? The hippocampus could be a possible target in this approach. - Mark Bolding, Ph.D., UAB, described brain neuro-modulation using transcranial focused ultrasound.
This high-intensity focused ultrasound is administered from outside the skull, and it can be guided with an MRI, Bolding, said. It offers high resolution (1 to 3 mm3) and is non-invasive. The ultrasound can target specific areas with variable intensities that can either heat or ablate that small part of the brain.
Focal ablation can be an outpatient procedure to treat essential tremor. Low-power ultrasound can modulate neuron firing, up or down. Local neuro-modulation is being tested in an epilepsy model.
The transcranial ultrasound can be used to give feedback from motor system prosthetics, or it can be used in functional mapping of the brain, as a complement to fMRI. The ultrasound can stimulate induced function or temporarily interrupt activity, Bolding said. It can help determine causality in networks, and the introduction of iron oxide nanoparticles that are ruptured by ultrasound provides a contrast agent for MRI.
“Transcranial focused ultrasound has a lot of future here at UAB,” Bolding said.
- Harrison Walker, M.D., UAB, said that more than 150,000 patients worldwide have received deep brain stimulators. He showed a video from a UAB surgery to show how nerve recordings during an operation to implant a pulse generator helped determine the best location for the pulses.
U.S. Food and Drug Administration-approved targets for deep brain stimulation are Parkinson’s disease, generalized dystonia, essential tremor and refractory OCD. Walker and UAB patient Bennie Burton also showed the audience how Burton’s Parkinson’s tremors return when Walker temporarily turns off his deep brain stimulation (for an example, see this 2013 UAB video of Burton).
Deep brain stimulation might help in depression, Walker said, but doctors cannot see an immediate effect like they do with conscious movement disorder patients during surgery, making it harder to know where to place the electrodes. - Other UAB speakers were Rajesh Kana, Ph.D., who talked of recent research to distinguish autism participants from control volunteers through multimodal neuroimaging; and Corey Shum, who described the use of virtual reality to simulate combat casualty care for U.S. Air Force Pararescue medics.
Amie McLain, M.D., UAB, the kickoff speaker, set the stage for the symposium by describing over a century of progress in putting a person’s life back together after incapacitating injury or disease. “We have now reached the technical age of rehabilitation enhancement,” McLain said. “With brain plasticity, we get the patient to do something to create new pathways in the brain, and this involves real recuperation of the brain … One of the biggest challenges now is the brain-machine interface, or the brain-computer interface.”
Such brain-machine interface research is inherently interdisciplinary and will require a different type of funding than the traditional NIH RO1 grants, Palazzo said. He noted that the issue of design —another interdisciplinary area — kept coming up in the talks.
“So, what can we do here at UAB?” he asked the participants. “We have some areas of strength already. We need to identify the gaps and identify areas where we might contribute.”
“Finally,” Palazzo said, “we will need to identify partners — other researchers, corporate partners, federal agencies and philanthropic partners. Progress can be made, but probably not by just one department, not by just one school, and probably not by just one university.”
The 2015 Bevill Neuroscience Symposium: Brain-Machine Interfaces
- Met on Feb. 27, at the UAB Alumni House.
- Honors former U.S. Congressman Tom Bevill, through the support of the Bevill Family Foundation and members of the Bevill family.
- Was presented by the UAB Comprehensive Neuroscience Center, the Department of Neurology in the School of Medicine, the College of Arts and Sciences and the School of Engineering.
- Ultimate goal of the Bevill Symposium in Neuroscience is to further our understanding of brain disorders and to develop new, improved treatments and cures.
Visiting Speakers
- Leigh R. Hochberg, M.D., Ph.D., associate professor, School of Engineering and the Institute for Brain Science, Brown University; vascular and critical care neurologist at Massachusetts General Hospital; and senior lecturer on neurology at Harvard Medical School.
- P. Hunter Peckham, Ph.D., Donnell Institute Professor and director of the Functional Electrical Stimulation Center, Department of Biomedical Engineering, Case Western Reserve University.
- Andrew Schwartz, Ph.D., professor of Neurobiology, University of Pittsburgh.
UAB Speakers
- Mark Bolding, Ph.D., assistant professor, Division of Advanced Medical Imaging Research, Department of Radiology
- Rajesh Kana, Ph.D., associate professor, Department of Psychology.
- Adrienne Lahti, M.D., the Patrick H. Linton Professor and director of the Division of Behavioral Neurobiology, Department of Psychiatry
- Amie McLain, M.D., chair and professor, Department of Physical Medicine and Rehabilitation
- Corey Shum, technical director, Enabling Technology Laboratory, School of Engineering
- Harrison Walker, M.D., associate professor and medical director for surgical movement disorders, Department of Neurology