19 June 2008—For more than a decade, doctors have been implanting devices called deep-brain stimulators into patients with Parkinson’s disease and stimulating a small area of their brains with low-voltage electrical pulses. So far, there’s been only one way to tell how patients are taking to the treatment: by watching. Are they walking smoothly again? Can they hold their hands in front of them without trembling? But a better way to evaluate treatment is to ask the brain directly. In such a system, neuronal feedback would direct the timing, location, and intensity of subsequent stimulation and would theoretically suppress side effects that many patients suffer today. A group of neural engineers from the University of Michigan, tackling a pivotal piece of the problem, have designed a programmable device capable of stimulating and recording from the brain simultaneously.
”It’s what a lot of people have talked about for a really long time, but nobody’s done it,” says Jerrold Vitek, a neurologist at the Cleveland Clinic, in Ohio, who treats patients with deep-brain stimulators and was not involved with the research.
Vitek describes the available technology as ”first generation.” The devices, manufactured by Medtronic, electrically stimulate the subthalamic nucleus, a structure deep inside the brain, through four electrodes. When electrical impulses hit the targeted cells, the tremors associated with Parkinson’s disease subside; however, the quality of treatment greatly depends on how well surgeons implant these four electrodes. A misplaced lead could stimulate surrounding tissue and cause changes in the patient’s mood and cognition. Such a positioning error was recently found to be a leading cause of the therapy’s failure. Even with a perfect implant, patients have only one control parameter: on or off. So what you have is an inflexible system trying to control a highly variable and plastic organ.
”The idea is that one size does not fit all in terms of stimulation programming,” says Daryl Kipke, who, with his colleagues, present their closed-loop system today at the IEEE-cosponsored Symposia on VLSI Technology and Circuits, in Hawaii. ”One way to provide more-specialized stimulation, or specialized treatment, through deep-brain stimulation is to develop a closed-loop system.”
Kipke and his colleagues have designed a system that integrates neural recordings from eight electrodes and uses them to program the amplitude, repetition rate, and duration of pulse generation in a 64-channel stimulator. Their goal is to get every component of the design onto a single chip, including the amplifier that connects to the probe, the data circuits, the digital filters, and the microprocessor that decides if, how, and when to stimulate. For now, however, the microprocessor is separate from the rest of the system, which resides on its own chip. ”A microprocessor gives information to the chip about where and how, and the chip takes care of the rest,” says Michael Flynn, a leader of the project.
In theory, a device like this could automatically detect side effects or poor performance and change the pattern of stimulation. But the obvious question is, what recordings will give you this information? Where are you going to put these eight electrodes, and what are they going to tell you?
It’s a question that no one in the field is ready to answer because, as Vitek points out, we don’t even really know how deep-brain stimulation works when everything goes right. In the 1980s, promising drug treatments proved to be only temporarily effective with tremor patients, which pushed researchers in the field to find new options. After observing that surgical brain lesions could quell movement disorders, doctors began inserting electrodes into the brain. One technique took chunks of brain offline completely; another stimulated them. But both seemed to work.
Since then, the picture has not gotten a whole lot clearer. But, says Vitek, ”if you wait until you’ve got everything figured out, you’re going to be waiting a long time.”
For now, Kipke’s device will serve as a research tool. ”This technology will help support research in animals that will lead to a better idea of signals that should be analyzed,” he explains.
The group is also trying to make the system more energy efficient. They claim to have already reduced the power consumption and size compared with other stimulators, characteristics that would translate into huge benefits for a clinical model.
About the Author
Morgen E. Peck is a freelance writer and former neuroscience researcher. In May 2008, she wrote about a robot that controls a brain-computer interface.