On the heels of one failed drug trial after another, a recent study suggests people with early Alzheimer’s disease could reap modest benefits from a device that uses magnetic fields to produce small electric currents in the brain.

Alzheimer’s is a degenerative brain disorder that afflicts more than 46 million people worldwide. At present there are no treatments that stop or slow its progression, although several approved drugs offer temporary relief from memory loss and other cognitive symptoms by preventing the breakdown of chemical messengers among nerve cells.

The new study tested a regimen that combines computerized cognitive training with a procedure known as repetitive transcranial magnetic stimulation (rTMS). The U.S. Food and Drug Administration has cleared rTMS devices for some migraine sufferers as well as for people with depression who have not responded to antidepressant medications.

Last month at the 13th International Conference on Alzheimer’s and Parkinson’s Diseases in Vienna, Israel-based Neuronix reported results of a phase III clinical trial of its therapy system, known as neuroAD, in Alzheimer’s patients. More than 99 percent of Alzheimer’s drug trials have failed. The last time a phase III trial for a wholly new treatment succeeded (not just a combination of two already approved drugs) was about 15 years ago. The recent study did not test a drug but rather a device, which usually has an easier time gaining FDA clearance. NeuroAD has been approved for use in Europe and the U.K., where six weeks of therapy costs about $6,700. The system is not commercially available in the U.S., but based on the latest results the company submitted an application for FDA clearance last fall.

The neuroAD setup resembles a dental chair fitted with a touch screen and flexible arms, which generate magnetic fields from metal coils positioned near the person’s scalp. The magnetic fields produce electric currents within the brain that influence the activity of neurons. The procedure can reportedly speed up learning by strengthening synaptic connections between neurons while the person performs tasks that engage those particular brain cells. In the cognitive training that accompanies rTMS, when study participants see a picture of a strawberry and touch the screen to identify it as “fruit” or “furniture,” for instance, the system stimulates Wernicke’s area, the brain region responsible for language comprehension.

For its latest rTMS trial, the company enrolled about 130 people with mild to moderate Alzheimer’s at 10 sites—nine in the U.S. and one in Israel. Four out of five participants were already taking symptom-relieving therapies. At the start of the trial, each person took a cognitive battery—a 30-minute paper-and-pencil test commonly used to gauge mental function in Alzheimer’s studies—and was randomly assigned to receive the rTMS-cognitive therapy or a sham treatment for six weeks. The sessions lasted about an hour each day, five days per week.

A week after the six-week regimen, and again five weeks later, participants retook the paper-and-pencil test to see if their cognition improved. Despite the elaborate protocol, study adherence was high. More than 90 percent of participants completed at least 90 percent of their visits, says Babak Tousi, who heads the Clinical Trials Program at Cleveland Clinic Lou Ruvo Center for Brain Health and reported the trial’s results at the Vienna meeting.

Based on past studies of the neuroAD system in smaller groups (none had more than 30 participants), the company expected to see a cognitive benefit after six weeks of treatment. Curiously, though, the recent study revealed no significant difference in test scores between active and sham groups at the seven-week time point. (The sham group sat in the chair and saw pictures on the screen but received no cognitive training or exposure to magnetic fields.) At week 12—six weeks after the therapy ended—the active group did show an 1.8-point test score advantage over the sham group. “That is a pretty small effect,” says Lon Schneider, who directs the State of California Alzheimer’s Disease Center and heard the study results presented in Vienna. By comparison, he says, drugs currently approved to treat Alzheimer’s symptoms have shown a 2.5- to 3-point improvement in six-month clinical trials. And in a study reported last fall, a leading pharmaceutical candidate tested in more than 2,100 people seemed to work about as well (a roughly 1.5-point improvement) but failed to achieve statistical significance.

Plus, the modest effect seen with the new rTMS trial only turned up in participants with mild Alzheimer’s, Tousi reported. People with more advanced cases did not improve on the therapy. “We’ve got that typical problem of a small study that does seem to give outcomes, but the outcomes are either unclear or not fully evaluable,” Schneider says, adding it is unclear, for instance, if the test scores improved because of the cognitive training or resulted from possible mood-enhancing effects of the rTMS, because some Alzheimer’s patients have depression or other psychiatric symptoms.

John-Paul Taylor, a neuropsychiatrist at Newcastle University in England who was not involved with the study and researches TMS’s prospects for treating visual hallucinations in dementia, agrees that it is hard to tell if the cognitive improvement was indeed “a real TMS effect.” He says, however, this technology is “ripe for more investigation.”

Taylor is working with colleagues who are trying to use computational modeling to get a better idea how rTMS works. “That’s where it’s going to get really interesting,” he says. “I suspect you’ll have to tailor the stimulation to individual patients.” Consistent with that idea, earlier this year researchers reported using brain imaging to identify different types of depression—and patients in one of those subgroups responded especially well to rTMS.

With the computational modeling, one could imagine feeding in a person’s brain scan “and the computer would say, you need to be in this position at this stimulation intensity to equal what another person would receive,” Taylor says. “That’s not that far off.” Ultimately, though, “we want a therapeutic that still works across everybody to some degree,” he says. “There’s a hint of that in this trial. I’m cautiously optimistic.”