Advancing Basic Science for Humanity
Scientists Want to Use Brain Implants to Tune the Mind
EPILEPSY, DEPRESSION, ALZHEIMER’S, post-traumatic stress disorder (PTSD)—what if the eventual treatment for these different brain conditions was not a pill or talk therapy, but some kind of implant?
Deep-brain stimulation (DBS) is already used to subdue the shakes and tremors of people with Parkinson’s disease. Electrodes are implanted into a specific part of the brain, connected via wires under the skin to a pacemaker-like stimulator in the chest. That pacemaker sends out electrical signals that stifle the parts of the brain that are causing tremors. Researchers are beginning to test whether similar devices, or new types of implants, could help people with other complex neurological conditions.
At the same time, a handful of projects devoted to creating the next generation of brain implants are being funded by the U.S. Brain Research Through Advancing Innovative Neurotechnologies (BRAIN) Initiative, a multimillion dollar effort to accelerate humankind’s understanding of the brain. One, Restoring Active Memory (RAM), aims to use implants to improve soldier’s memories after traumatic brain injury. Another, called the Systems Based Neurotechnology for Emerging Therapies (SUBNETS) program, is developing devices to treat PTSD, chronic pain and anxiety.
But these devices face many hurdles en route to the clinic. For any given disorder, what nerve cells are the problem, and how can stimulation or a brain chip set them aright? What materials will work best with brain tissue? Can implants be made wireless and small enough to fit into a skull?
The Kavli Foundation spoke with three scientists—two of whom recently participated in a discussion on the clinical implications of the BRAIN Initiative—about where brain implants are today and where they are headed.
The participants were:
- HELEN MAYBERG, MD–is the Dorothy C. Fugua Chair of Psychiatric Neuroimaging and Therapeutics and professor of psychiatry, neurology, and radiology at the Emory University School of Medicine in Atlanta. She has been studying the use of deep brain stimulation to treat depression for more than a decade.
- BRIAN LITT, MD–is Director of the Center for Neuroengineering and Therapeutics at the University of Pennsylvania. Litt is creating flexible electrode arrays and implantable devices, and also uses cloud computing, to identify and suppress seizures in people with epilepsy.
- JOSE CARMENA, PhD–is Chancellor’s Professor of Electrical Engineering and Neuroscience at the University of California-Berkeley, and co-director of the Center for Neural Engineering and Prostheses at UC-Berkeley and the University of California-San Francisco. Carmena has co-developed implantable microsensors and is an investigator on the SUBNETS project.
The following is an edited transcript of their roundtable discussion. The participants have been provided the opportunity to amend or edit their remarks.
THE KAVLI FOUNDATION: Helen, not long ago, using a brain implant to treat a disease such as Parkinson’s was seen by some as radical. Today, it’s not only accepted; scientists like yourself are looking to expand the possible applications of this therapy. What’s changed?
Helen Mayberg has been studying the use of deep brain stimulation to treat depression for more than a decade. She is a neurologist and researcher at the Emory University School of Medicine in Atlanta. (Credit: H. Mayberg)
HELEN MAYBERG: Well, it wasn’t too radical. It was a logical application of everything we knew about neuroscience.
We’ve come a long way in terms of how we understand diseases of the brain. Before, we thought of them as the result of a localized abnormality in one part of the brain. But it isn’t one location and one function gone wrong and causing disease. Instead, within the brain are many regions that work together in modules or circuits, in many ways as if choreographed in a dance. For example, there are several brain areas working together to produce movement or speech. When any one of those areas is altered or damaged, it affects the whole network. Therefore, we now think of many conditions as “neurocircuit disorders.”
JOSE CARMENA: In terms of technology, the know-how we have acquired in the last 10 to 15 years is also beneficial. For example, in my group, through the ongoing development of brain-machine interfaces, we have learned a lot about how to target different areas in the brain. The aim of these devices is to help people who are paralyzed by injury or disease move again by creating an artificial pathway between the areas of the brain that control motion and the muscles. Essentially, they are devices that use electrodes implanted in the brain to translate thoughts into the action of prosthetics.
BRIAN LITT: Another exciting change is that techniques like deep brain stimulation, which uses electricity to control nerve cells, have improved so that we can now target specific kinds of cells in specific places. This lets us better focus the treatment on the part of the brain that needs help. This is markedly different from traditional medications, which affect the whole brain at one time.
MAYBERG: We basically have a new way to think about these disorders, new tools to study the brain and technological advances to affect its activity. It’s really a very unique time when knowledge and technology are interacting.
TKF: The U.S. government, through the BRAIN Initiative, has funded a handful of projects to develop new technologies that can change the way the brain functions. Is this one of the game changers?
Brian Litt is creating flexible electrode arrays to identify and stifle seizures in people with epilepsy. He is a neurologist and professor bioengineering at the Perelman School of Medicine at the University of Pennsylvania. (Credit: UPenn)
LITT: The BRAIN Initiative was revolutionary in taking a large chunk of money and saying, “We’re going to aim to develop new technologies, to be able to map and control networks of neurons in the brain, in order to treat disease and change behavior.”
The technology that’s in many of the implantable devices we use in patients today has not changed much in recent years. We’ve been putting platinum electrodes into brain tissue and connecting them to a box under the collarbone that stimulates tissue. Now, there are people working on all kinds of new devices, which are capable of things such as controlling neurons with light, or ablating, listening to or controlling them with ultrasound. Some groups are even making electrodes that grow themselves, or can be moved after they have been implanted in the brain. This is a dramatic change in the landscape.
CARMENA: I agree. The BRAIN Initiative has enabled the entire community to think big and to imagine the future.
TKF: At a scientific meeting last year, Helen, you and many others discussed the potential use of deep brain stimulation to treat neurocircuit disorders ranging from Tourette syndrome to Alzheimer’s to depression. How do you know if electrical stimulation will be successful for a certain condition?
MAYBERG: That’s the million dollar question. It’s not as simple as copying what is done in Parkinson’s disease. For example, a Parkinson’s device stimulates the brain all the time, but to treat something like epilepsy, you only need occasional stimulation to block a seizure.
LITT: First you have to figure out the nuts and bolts of the circuit causing the disease. That will indicate whether you need to enhance or suppress function, and in what particular area of the brain. Then you have to ask, “How invasive is the technology? How well is it tolerated? What are the side effects?” You can do a significant amount of that research in animals, but an animal isn’t going to be able to tell you if they get numbness in their face every time the stimulator goes on, like a person with Parkinson’s would in the operating room. So at some point you take that leap to try it in people, and we’re beginning to see that.
TKF: Speaking of Parkinson’s disease, have we reached the limit of what implants can do for that condition, or can the therapy still be improved?
Jose Carmena has developed brain-machine interfaces and implantable microsensors. He is a professor of electrical engineering and neuroscience at the University of California-Berkeley. (Credit: UC-Berkeley Electrical Engineering and Computer Sciences)
LITT: There’s a lot that’s still going on for Parkinson’s. People are trying to stimulate more and new targets in the brain. They are also working on more superficial forms of brain stimulation, which may be more effective or augment current methods. And, finally, they are manipulating brain signals in different ways. I think having more intelligent devices that read the patient’s brain activity and automatically adjust their output to that activity will also improve treatment. Helen, what about you?
MAYBERG: There’s room to build a better mousetrap.
Keep in mind, while stimulators help a lot of people with Parkinson’s, they only treat tremor and rigidity. People continue to fall, they have difficulty thinking, they have wild fluctuations in their behavior. The implants don’t treat those other symptoms. Leveraging what scientists learn in studies of implants for anxiety or depression or memory may have implications for how to treat these other Parkinson’s symptoms.
MAYBERG: For depression, we target a different part of the brain. We’ve had good results targeting the subcallosal cingulate and adjacent white matter. This approach impacts not only this region of the cerebral cortex that regulates mood, but also its connections to other parts of a likely depression circuit throughout the brain.
One challenge has been deciding who to test the implants on. When is depression so severe that you would even think to do an implant? When do the varied symptoms become so disabling, resistant to treatments like medications or therapy, or likely to lead to suicide? You have to balance the potential benefit and the potential risk.
This is why much of the work I’ve done on humans is with the most severe and intractable cases of depression. Perhaps if the treatment is effective for the most severe cases, we can consider patients who are less severe, to help them avoid relapses or worsening of their depression. We don’t know where those risk/benefit boundaries are, so we have to do our best to test the therapy safely and reasonably, in patients who have reached the stage where they have limited other options.
To treat depression, Mayberg and colleagues target their electrodes to the subcallosal cingulate and adjacent white matter tracts, a brain circuit involved in regulating mood. (Credit: H. Mayberg)
CARMENA: Over the years, I’ve seen more and more studies on implants and memory. Many of them are being pushed by aggressive research agendas, coming from the Defense Advanced Research Projects Agency (DARPA) and other federal agencies involved in the BRAIN Initiative.
LITT: These are very early days, but we’ve achieved very small improvements in memory so far—sort of like remembering an extra couple of words on a list. Kathryn Davis from our center is involved in a trial, here at the University of Pennsylvania, led by Michael Kahana: the $20-million RAM project to enhance memory, sponsored by DARPA. The idea is to map out brain circuits that are involved in memory, and do selective brain stimulation to see if we can enhance it. At centers around the country, the team has started to map those circuits in epilepsy patients who have electrodes implanted in their brains to locate the sources of their seizures prior to surgery.
There are also several start-up companies doing similar research. One of them, Kernel, is based on technology developed by an investigator named Ted Berger at the University of Southern California in Los Angeles. He’s been doing brain stimulation in rodent models and trying to work out how to enhance memory.
MAYBERG: The answer is yes. For example, we’re gaining a better understanding of the brain circuits involved in depression by combining our results in patients with implants with brain imaging data.
LITT: Implants have already taught us a lot about how the brain is built, and how it functions in healthy people and also in disease states. For example, short-term implants have taught us a lot about epilepsy and how to treat it. Surgeons used to remove a lot of brain tissue because people thought that epileptic seizures started over a large region of the brain at once. Now, thanks to recordings from implants as well as external electrodes, we can often narrow the source to smaller regions. Then, we can use a laser to ablate that smaller part of the brain to prevent seizures and greatly improve a person’s quality of life.
To record more closely from the nerve cells that instigate an epileptic seizure, Brian Litt and colleagues have developed flexible electrode arrays. (Credit: Travis Ross and Yun Soung Kim, ITG Visualization Lab, Beckman Institute, University of Illinois)
Long-term implants allow us to record from the brain continuously to understand what happens during sleep and during wakefulness, and how much variation and fluctuation in brain activity there is in disease. With long-term recording, we can even address questions like how the brain changes during aging because we have what amounts to a diary of brain activity.
LITT: For the implants themselves, we need electrodes that are more compatible with biological tissue and we need batteries that last much longer. We also need systems that can do more computations on less energy and that can hold and transmit more data.
Beyond the implants, there’s a lot of room for innovation in technologies that can identify the specific groups of nerve cells that we want to stimulate. You see, electrical stimulation kind of blasts a whole area. Ideally, you’d like to select specific neurons, so that you won’t cause a lot of side effects.
CARMENA: I agree. I think it’s amazing that electrical stimulation works as well as it does given its limited spatial resolution. Still, there’s a lot of room there for better specificity. The most obvious way of achieving that is through optical stimulation, such as optogenetics, which uses light to stimulate specific cell types. The problem is that you have to inject a virus to genetically modify the cells you want to respond to the light. But I believe there will eventually be virus-free optical techniques to be used in humans.
There’s also a lot of computer programming that goes into this science. For example, we have to model how the brain works and how our devices will affect it, and design “artificially intelligent” devices that can handle the computations inside a patient rather than remotely. What I’d like to see, eventually, is in one single implantable device with the capacity to identify the problem nerve cells, figure out what pattern of stimulation will fix them, and apply that stimulation.
TKF: Deep-brain stimulation isn’t the only kind of neural implant under development. Jose, you’ve engineered an implant to activate multiple areas of the brain at once. Could you tell us about your “Neural Dust” sensors and what they can do?
CARMENA: With Michel Maharbiz and other Berkeley colleagues, we developed Neural Dust, an ultrasonic backscattering system to record information from the human body. This was motivated by one of the main problems in brain-machine interfaces: how to make a technology that will last in the brain for decades, if not a lifetime, and that is also wireless. Neural Dust uses ultrasound waves to power tiny, wireless sensors that can be implanted in the body and read its output.
UC Berkeley bioengineers Jose Carmena and Michel Maharbiz invented tiny sensors they call “neural dust” to wirelessly monitor, and even stimulate, the brain and other organs. (Credit: UC Berkeley)
Think of it like a Fitbit, but internal.
The main part of the sensor is a piezocrystal, a material that vibrates at a given frequency. When you apply ultrasound from outside the skull, the sound vibrates the crystal, and the energy of that vibration powers a simple electrical circuit and a sensor. The sensed electrical activity of nearby nerve cells changes the electrical impedance which in turn changes the vibration pattern of the crystal. As a result, the echo coming back from the “dust” mote contains information from the electrical activity of those cells.
We’ve used Neural Dust to make the very first recordings of nerve and muscle activity from living rats. Eventually, we believe Neural Dust could be used to monitor organ function throughout the body and influence brain activity, for example by preventing seizures.
TKF: There are other ways to stimulate neurons without surgery or implants at all. Do you envision a day when the field has moved beyond implants and adopted devices that are located outside the brain?
LITT: It would be wonderful if we didn’t have to do surgery to get devices into place. There are people, in our group and others, working on how we might inject materials into the blood that would somehow get into the brain and self-organize into devices.
MAYBERG: I’m hoping that what we do now will become obsolete, and hopefully sooner rather than later. I imagine that 10 years from now, we will be fine-tuning explicit brain circuits without invasive implants.
LITT: We work on practical ideas in the short term, but dream about bigger things in the long term. I think we’re only limited by our imaginations.
—Amber Dance, Winter 2017