<b>Neuroscience: Still Getting Wired </b><br/><br/>

Neuroscience: Still Getting Wired

The simple electrode, long an unglamorous workhorse of neuroscience, may become the field's star tool in the next few years.

Figure 1. An experimental microwire-based multi-electrode array for chronic implantation in a primate brain. Image courtesy Miguel Nicolelis, Duke University.

Mentioning neuroscience to most people these days generates a recursive loop: their minds picture a picture of a mind. The visually arresting images of modern MRI and PET scans have become the field's de facto brand, making it easy to forget the much humbler tools and techniques that under-gird much of our current understanding of the brain.

Indeed, some of the most exciting advances in neuroscience in the near future may have nothing to do with full-color scans of whole brains, and everything to do with a much older method: attaching electrodes to a neuron and measuring the changes in its electrical potential over time. While more graphically interesting experiments continue to bask in the limelight, experimentalists and equipment makers have engineered a quiet revolution in electrode design, opening the door to an astonishing new wave of basic science and clinical breakthroughs.

Probing longer

Like nearly every electronic device, neuroscientists' electrodes have gotten progressively smaller, more efficient, and more densely packed as they have evolved. While the first electrical studies of neurons focused on isolated cells in petri dishes, researchers have repeatedly tried probing electrical potentials inside living brains. In the 1980s, the wires finally got small enough to do that without killing the animal. "The initial technology was just wires, which are really good laboratory tools because they're inexpensive," says John Donoghue, Ph.D., Chief Scientific Officer and founder of Cyberkinetics (Foxborough, MA).

Figure2. Miguel Nicolelis, MD, PhD, is developing brain-machine interfaces in primates, using microwire-based multi-electrode arrays. Images courtesy Duke University.

From single microwires, researchers soon moved up to multi-electrode arrays, and manufacturers now offer a range of systems that can probe tens or even hundreds of neurons simultaneously, feeding the output directly to a computer for analysis. "This thing exploded about 10 years ago... because of the interest in clinical applications in addition to the basic science work, so there are many types of electrode arrays these days," says Miguel Nicolelis, Ph.D., professor of neurobiology at Duke University Medical Center (Durham, NC).

Cyberkinetics, for example, has developed what Donoghue characterizes as "second-generation" arrays, which pack as many as 100 probes into a square 4 mm on a side. The probes are insulated along their length to prevent short circuits, while their bare points contact individual cells within a brain region.

Initially, researchers implanted the arrays acutely, looking at changes in neuronal activity over short periods in anesthetized animals. The field's cutting edge, however, is now aimed at chronic implantation, leaving the arrays installed for months and monitoring cellular-level signals as animals engage in their normal behavior. While much more interesting scientifically, these chronic implantation studies raise a host of thorny technical problems.

Figure 3. Miguel Nicolelis, MD, PhD, is developing brain-machine interfaces in primates, using microwire-based multi-electrode arrays. Images courtesy Duke University.

"If you use some of the designs that have been proposed out there you create a mega lesion and reaction in the brain, and within weeks your recordings disappear," says Nicolelis. Shrinking the probes and using the thinnest possible wires helps reduce the inflammatory response, but Nicolelis warns that this can cause a different kind of trouble. "If it's too small, too thin, the deposition of the extracellular material [creates] an interface of gunk there that doesn't allow you to record signals from cells," he says.

While Nicolelis relies on microwire probes, other scientists prefer silicon-based systems, especially the family of technologies known as Michigan probes. Borrowing fabrication technology from the semiconductor industry, researchers at the University of Michigan first developed these highly reproducible probes in the 1970s.⁽¹⁾ NeuroNexus Technologies (Ann Arbor, MI) now markets a widely used line of Michigan probes for acute use in animals, and scientists at the university are adapting the technology for chronic implantation as well.

This is your brain online

Whether they are implanting multi-electrode arrays for a few minutes or a few months, neuroscientists are finding the technique extremely useful. Electroencephalograms and imaging methods can record gross electrical and circulatory changes around the brain, but only electrodes can tap into individual axon potentials. "This sensitivity gives us direct access to the communication language of the brain for multiple neurons, so we can for the first time study multiple neurons as they interact and function," says Donoghue.

Figure 3. Artist’s rendering of a recent experiment, in which a monkey controlled a computer game with a joystick, and then with only a multi-electrode-based brain-machine interface. Image courtesy Miguel Nicolelis, Duke University.

Nicolelis concurs. "This has become a real premiere paradigm to study very basic questions in brain research, like trying to clarify fundamental issues of how brain circuits represent basic information," he says. That information can range from the signals that underlie intention and movement to the neuron-level changes that accompany pathological processes such as seizures.

Many of the recent basic science breakthroughs with this technique have pointed the way to obvious clinical applications, some of which border on science fiction. Nicolelis's group, for example, is perfecting methods for hard-wiring a brain to a computer. By implanting an array in a motor region of a monkey's brain, the investigators were able to monitor the signals driving arm movement while the animal played a simple video game with a joystick. Feeding these signals into a powerful computer cluster, they eventually developed a digital model of the underlying control language; in computer terms, they had decoded the hardware driver for a monkey arm.

In the next phase of the work, the team took the monkey's joystick away, and handed it to a robotic arm that was connected to the computer. After a short learning period, the monkey started playing the game by controlling the robot arm directly with its mind, without moving its biological arm.

Conveniently, evolution seems to have conserved the arm-driving language in primates. In a recent clinical test, the researchers found that humans can also control the robotic arm mentally. "This is allowing people to design completely new experiments to... use the output of the brain to control devices that are very different from the biological activators that the brain normally controls," says Nicolelis. The same strategy seems to work with lower limbs, and the investigators are now testing a system that could allow a primate brain to control a pair of robotic legs.

Figure 4. Electron micrograph of a microwire-based multielectrode array optimized for chronic clinical implantation. Image courtesy Cyberkinetics.

Clinical applications are also a major emphasis for Cyberkinetics, which ran the first successful trial with multi-electrode arrays implanted chronically in humans. The company hopes to develop a practical system that will allow paralyzed patients to regain mobility, either by controlling robotic limbs or by feeding movement signals from the brain to the patient's actual limb, effectively bypassing the damaged nerve segment.

Wiring the future

Even while pointing to the technique's enormous research and clinical potential, experts caution that chronically implanted multi-electrode arrays are far from mature. Scientists can buy a range of appropriate arrays for acute use, but the exacting requirements of chronically implanted arrays have kept them out of mass production. While researchers like Nicolelis often fabricate their own arrays in dedicated facilities, smaller groups can turn to facilities like the Center for Neural Communication Technology (Ann Arbor, MI).

"Our service component involves working with collaborators around the country [and] developing devices that would allow the collaborators to do more advanced experiments," says Daryl Kipke, PhD, the center's director. Building on the successful Michigan probe design, Kipke and his colleagues now fabricate increasingly complex electrode arrays for chronic implantation.

Figure 5. Diagram of a typical human trial, interfacing an implanted multielectrode array with a desktop computer. Image courtesy Cyberkinetics.

The investigators are also adding new features to the system. "Now we're into polymer devices, fluidic devices, and composite structures," says Kipke. One major focus for the Michigan team is to add chemical sensors and fluid delivery systems to the arrays, allowing them to sense both electrical and chemical activity in neuronal clusters, and eventually deliver doses of drugs or neurotransmitters to change the cells' responses. "We're at the proof-of-concept stage there, but the outlook is very good," says Kipke.

Allowing the array to interact with the brain chemically and electrically will let researchers not only monitor neuronal activity at high resolution, but also change it with single-cell precision. In the clinic, that could enable a whole new range of therapies for neurological and even psychiatric diseases. "You can imagine one day, a closed-loop system, and there might be a little drug pump that will deliver the correct kind of medication at the correct site," says Donoghue.

Looking even further ahead, decoding the low-level language of the brain might eventually allow neuroscientists to move the sensing probes to the outside of the skull, and read the same signals non-invasively. "If that appears, then you're talking about a true revolution, because then a healthy human being would be able to take advantage of this technology," says Nicolelis, adding that among other things, "that would change completely the way we interface ourselves with our computers."

Figure 6. Artist’s rendering of a proposed microelectrode array system, with a wireless connection outside the patient’s head. Image courtesy Cyberkinetics.

Though antidepressant implants and keyboard-less computing may be decades away, Donoghue projects that at least a few severely paralyzed patients will be benefiting from multi-electrode-based systems within five years. He adds that "in terms of experimental work, there's going to be an explosion of labs studying large scale phenomena in the brain things you can't get out of MRI."

1. Wise, K.D. et al. An integrated-circuit approach to extracellular microelectrodes, IEEE Trans Biomed Eng. 17(3):238-47 (1970).