That psychiatrists can use both drugs and electricity to battle illness testifies to the fact that the brain is both a chemical and an electrical organ. Every brain cell has a halo of short projections attached to its body and a long trunk, called an axon. To communicate with another cell, it sends a pulse of voltage down the axon. The axon usually terminates at one of the short projections of another brain cell. Rather than make a direct electrical connection, two brain cells communicate via a puff of chemical transmitters released from the end of the axon when the voltage pulse reaches it. These transmitters cross the nanometers between the end of the axon and the next cell's projections and bind with receptor molecules there. Depending on the type of chemical signal, this binding can lead to a variety of things, but the simplest is an influx or outflow of current that briefly raises or drops the target cell's voltage. The cell integrates the voltage changes from its many projections, and, if the combination of them is big enough, it will trigger a voltage pulse down the target cell's axon. The process of integration and signaling continues as signals propagate through the brain's millions of specialized circuits and is the basis of everything that occurs inside our heads: thoughts, emotions, moods, memories, and dreams.

Psychoactive drugs, such as Prozac, work on the chemical side to ultimately affect electrical signals. Depression, at least in part, involves a problem with the electrical signaling between certain parts of the brain whose cells signal with a chemical transmitter called serotonin. By inhibiting the reabsorption of serotonin, Prozac lets more of the chemical accumulate in the space between the end of the axon and the next brain cell, thus restoring the signaling.

One problem with this approach is that drugs work everywhere in the brain that their chemical target exists, regardless of whether those parts have anything to do with depression or any other disease, and that leads to side effects. Prozac, for example, has been known to reduce sex drive and can cause insomnia. Another problem is that brain chemistry varies from person to person, so no single drug will work in everyone.

The shared goal behind the new electromagnetic therapies, on the other hand, is to use electricity itself to restore the signaling, ideally, only in those parts of the brain affected by disease. Decades ago, neuroscientists demonstrated that electrically stimulating a neuron alters, in the long term, the strength of its connections to other neurons—making an electrical signal from one neuron more likely or less likely to jump to the next neuron. Though little is known in detail about how the new therapies work, it's likely that, to varying degrees, they depend on that phenomenon.

Because they are new and in some cases relatively unproven, the device-based technologies are being tested exclusively in people for whom all the available drugs have failed to work. For a minority of these patients, electroconvulsive therapy, a 70-year-old technique, is the treatment of last resort. So it is with electroconvulsive therapy that the new technologies are generally compared.

Unfortunately, your view of electroconvulsive therapy, like that of many potential patients, was probably formed by the 1975 movie One Flew Over the Cuckoo's Nest, in which it was used as a means of punishment and control. Even if Jack Nicholson's performance has no influence on your view of psychiatry, the idea of the therapy's main side effect, amnesia, is far more fearsome than Prozac's decreased libido or even the maladies associated with more powerful drugs, because memory is so tied up with our sense of self. But the reality is that the severity of electroconvulsive therapy's side effects has been minimized over the years, its use is carefully controlled, and, quite simply, nothing is as effective at breaking through the worst forms of depression. Still, in the United States, only about 100 000 people a year agree to it, despite the millions whom no drug helps.

"Electroconvulsive therapy can be dramatically effective at restoring a person's health and getting their life back on track," says Sarah H. Lisanby, director of the Brain Stimulation and Neuromodulation Division of the Columbia University Medical Center, in New York City. "The potential for the new brain stimulation techniques is to get those kinds of dramatic effects in medication-resistant populations without the downside."

Vagus nerve stimulation began in the 1980s with Jacob Zabara, a neurophysiologist at Temple University, in Philadelphia, demonstrating that he could quell epileptic seizures in a dog by electrically jolting its vagus nerve, one of twelve pairs of nerves that emerge from the brain instead of the spinal cord. He showed the technique to pacemaker designer Reese Terry, and a few years later they formed a company called Cyberonics Inc., in Houston, to develop a treatment for epilepsy.

Using off-the-shelf integrated circuits, design help from friends in the field, and a new kind of helical electrode, Terry put together an implantable device that periodically shocks the vagus nerve. Cyberonics has made more than 30 000 of them, using the same basic design. The implantable device looks and acts like a heart pacemaker. Though a doctor can program in a wide range of stimuli, the device typically delivers 1- to 2-milliampere, 250-microsecond pulses at 20 to 30 hertz for 30 seconds every 5 minutes.

Terry and his co-workers always envisioned uses beyond epilepsy. Depression was a good place to start, because the malady has been linked to epilepsy for so long that even Hippocrates wrote about it. About a quarter of people with severe epilepsy also have chronic depression—a far greater ratio than in either the general population or other groups with chronic illnesses. Also, intriguingly, early in Cyberonics' tests, some epilepsy patients reported that the device had improved their mood.

Researchers don't really know why the device works against depression. But they do have some theories. Phillip C. Jobe at the University of Illinois College of Medicine, in Peoria, proposes that the brain's natural defenses against both epileptic seizures and depression are weakened by chemical and structural flaws in the same two systems of neurons buried deep within the most primitive part of the brain. Vagus nerve stimulation alters activity in both those areas, although the nerve does not connect directly to either of them.

Terry, naturally, takes an engineer's view of things. "The way I look at it," he says, "the brain is a very finely controlled feedback system." For some diseases, he suggests that the "control system is a little bit out of balance." The periodic pulses from his device in effect "pace" the vagus nerve, he believes, restabilizing the control system.

But a bigger question than how it works, and one the company is still trying to answer for doctors, is whether or not it actually does work. In the late 1990s, a pilot study of patients with chronic or recurrent depression that resisted treatment with drugs gave promising results. McGuffee was among the first patients to receive an implant, in February 1999. One month after she got the implant, her family began to see an improvement; a few months later, McGuffee noticed it, too.

The pilot study was enough to convince European and Canadian regulators to allow the stimulator's use in their jurisdictions. To get more conclusive data that might satisfy the tougher U.S. regulations, Cyberonics embarked on a 235-patient, eight-week study. To tease out any placebo effect, all the patients received implants, but only half of the implants were turned on. Here again, too few patients improved to tell if the device was the cause of the improvement. So at the end of the study the company asked doctors to turn on the implant for anyone who wanted it and instructed them to continue treating the patients with anything that might benefit them. "It would have been inappropriate to withhold treatment," says chief medical officer Richard Rudolph. "But now we had nothing to compare the outcomes with."

Strapped for cash but not ready to give up on a group of patients with no options, to say nothing of a potential US $1 billion market, the company continued to try to prove the stimulator would work for depression. Plan B, according to Rudolph, was to follow the patients from the original study, find a group of very similar patients without stimulators, and compare how they fared over two years, a much longer period than is generally used in a trial of a new antidepressant drug.

After one year, one in six patients treated with the nerve stimulator was free of depression, and 56 percent got some meaningful benefit—as measured by a standardized questionnaire used to rate the severity of a patient's depression. Of those who did respond, about 70 percent continued to benefit after two years. But waiting a year to see if the treatment worked in a disease that comes at irregular intervals was highly unusual. The lack of a control group that had the device implanted but not turned on to counteract the placebo effect was stranger still. In August 2004, the U.S. Food and Drug Administration, which regulates the marketing of medical devices, decided not to allow Cyberonics to sell the vagus nerve stimulator as a depression treatment, overruling its own advisors in the process.

Cyberonics' CEO, Robert P. ("Skip") Cummins, who lost both his mother and grandfather to depression-related suicide, refused to give up. His company gathered more data, and went straight to the FDA's top brass. By February 2005 the company had won conditional approval. But it still had hoops to jump through on the way to full approval: there was controversy when Public Citizen, a prominent Washington, D.C., advocacy group, questioned whether the device worked at all. At the same time, an investor lawsuit began regarding the timing of some executive stock sales. And then there was a halfhearted investigation by a U.S. Senate committee into why the FDA had decided against the device. Full approval finally came last July.

Cyberonics says it has trained 2000 psychiatrists in vagus nerve therapy so far, but many physicians are still skeptical. Perminder Sachdev, a professor of psychiatry at the University of New South Wales, Sydney, Australia, thinks the technology has shown some promise but has a way to go before the results are convincing. "It's a hard area to investigate," he says. The placebo effect is difficult to eliminate, the nature of depression is that it waxes and wanes, and the treatment takes a long time to show an effect. The combination of all that means you need a great many patients to prove a device is working, he believes. Sachdev and others expect the picture to clear somewhat after the results of a study going on now in Europe are reported. In the meantime Cyberonics is running pilot trials to see if the device will work to control other mental illnesses, such as bulimia and obsessive-compulsive disorder.