Martha Rhodes experienced her first bout of depression at 13. By her late 50s, she had taken just about every antidepressant there is, including Zoloft, Lexapro and Paxil — which did the trick for many years, but had side effects — then Effexor, Lamictal, Seroquel and Abilify.

After a suicide attempt in 2009, she tried something radically different: transcranial magnetic stimulation, or TMS, a treatment in which magnetic pulses are used to stimulate parts of the brain believed to be involved in mood regulation. Unlike electro convulsive or shock therapy, which is also used to treat stubborn depression, TMS does not generally produce seizures.

Every day, she spent just over half an hour in a chair with a powerful magnet affixed to the front left side of her head. After four weeks, "I woke up and something was different," said Rhodes, who wrote a book, "3,000 Pulses Later" describing the treatment. "I felt lighter. I didn't wake up in the morning and wish I were dead."

For Rhodes, 63, a former advertising executive in Danbury, Connecticut, TMS treatment was transformative, and she no longer needs antidepressants. But there are still many questions about just how many severely depressed patients respond to TMS, which requires daily office visits for several weeks, costs thousands of dollars and is often not covered by insurance.

For the therapy, patients sit in a doctor's office with a large magnet pressed to the left side of their heads. The idea is that a pulsed magnetic field, similar to that used in M.R.I.'s, creates an electrical current in the surface of the brain that "resets" the patient's mood regulatory system.

TMS is approved specifically for patients whose depression does not respond to antidepressants or who cannot tolerate the drugs' side effects, which include weight gain and loss of sex drive. Many of these patients are desperate for alternative treatments, but it's not certain that TMS can provide the help they need.

"While it's fairly clear that TMS is effective in some percentage of patients with major depressive disorder, it's still not very easy to know in advance who those patients are," said Dr. Steven J. Zalcman, the head of the clinical neuroscience research branch of the National Institute of Mental Health.

The American Psychiatric Association's practice guidelines say that TMS confers "relatively small to moderate benefits," and the results of clinical trials have been decidedly mixed. "The glass is half full or half empty, depending on where you're coming from," said Dr. Mark S. George, a professor at the Medical University of South Carolina in Charleston.

Dr. George was an investigator in the first independent, randomized trial of repetitive TMS to treat resistant depression. While that study found that nearly three times as many patients went into remission after TMS treatment, compared with those receiving a placebo, the absolute numbers were still small: Only 14 per cent of treated patients recovered, compared with 5 per cent in the placebo group.

A newer variation of the treatment, deep TMS, developed by an Israeli company called Brainsway and approved by the Food and Drug Administration this year, brought a higher response rate in a randomized trial: 30 per cent of 233 patients in the trial achieved remission, compared with 14.5 per cent of the control group.

The device targets deeper regions of the brain, like the nucleus accumbens, which plays a pivotal role in the reward circuit of the brain, company officials say.

An advantage of TMS treatment is that it is not invasive and, unlike medication, appears to have few side effects, just occasional discomfort or mild pain in the scalp at the site of the treatment, or headaches. Few patients drop out of treatment because of the pain, however.

Long–term effects are not known, but TMS does not take a toll on memory and cognition, as electroconvulsive therapy can. In rare cases it may cause seizures.