AS PART OF MY JOB reporting on neuroscience, I found myself in an unusual situation 10 years ago. During an interview, I offhandedly told a researcher to contact me if he ever needed a volunteer for a study. Months later the neurologist actually called, and I enrolled in a project on Parkinson's disease.

I was soon lying in a positron-emission tomography machine. Scientists injected a radioactive dye into my left arm, which felt warm and tingly as it coursed toward my brain. It settled in the regions that produce dopamine, a chemical that becomes depleted in Parkinson's. The researchers hoped that brain scans of middle-aged people could reveal the earliest signs of dopamine loss.

My dopamine levels turned out to be very high, “the highest we've seen in a normal volunteer,” the neurologist told me. But he and his colleagues had also found something unsuspected. They wanted me to undergo magnetic resonance imaging (MRI) to highlight one particular area. Alarmed, I agreed, and a few days later they took the MRI scan. There, in a dark corner of my cerebellum, was a large ghostly-white mass. It wasn't pretty. The researcher, who was not a physician, shrugged uncomfortably. The radiologist said nothing. Many hours later the neurologist called and told me the shadow indicated a cyst and not to worry. I had probably had it since I was born.

What “IF” Scenarios

No one volunteers for a study expecting that something sinister may appear. But now, after more than a decade of brain-imaging research, scientists have run across the likes of me often, and they have finally pulled together to discuss the issue of such “incidental findings”—IF, for short. The findings range from tumors and blood clots to cysts and other structural abnormalities. Investigators simply don't know what to do when they happen on these anomalies in what are supposed to be “normal” test subjects.

This past January dozens of scientists, lawyers, ethicists and policymakers convened at the National Institutes of Health to debate the issue. It seems that incidental findings show up in 20 percent of subjects in research studies—a huge number—and there are still no official procedures for handling such discoveries. Judy Illes, a senior research scholar at the Stanford Center for Biomedical Ethics who organized the meeting, notes that researchers typically are not medical doctors and shouldn't be put in the position of practicing medicine. Yet they become good at sizing up scans. What should they do when they spot something? When should they scan and tell?

The answer is not easy, Illes and other experts note. Scans are like Rorschach tests: in the best hands, scans can still be interpreted differently. No one even agrees what a normal brain should look like.

“It's a judgment call,” says David Eidelberg, director of neuroscience at the North Shore Long Island Jewish Health System in Manhasset, N.Y. “There are lots of variants of normal. Do you tell a person that they have a cyst in their brain that will never alter the course of their life? I'm not sure.”

This position, in essence, is that there is no reason to alarm a test subject unnecessarily. But others say volunteers should be told about any kind of unusual indications, whether it might worry them or not. It's their brain, and they should be informed about it. The sticking point is that scientists have no uniform way of handling incidental findings. Illes and her colleagues want to adopt a basic framework to follow. “The idea is to come up with solutions to protect our research volunteers, our patients and our institutions,” Illes says.

Call a Doctor

B. J. Casey, a psychologist at Weill Medical College of Cornell University who took part in the NIH meeting, recounted the first time that her team stared at an odd white mass on the MRI scan of a normal volunteer. “We all realized it was something that shouldn't be there,” Casey says. “But we aren't doctors, we aren't neuroradiologists, we're researchers.”

Casey did call in a neuroradiologist, who concluded that the mass was a brain tumor. “We saved this person's life,” she states. Nevertheless, the situation and others like it are uncomfortable. “We don't want to enter into a patient-doctor interaction in a research study,” Casey says. “We have to separate research from clinical practice.” Then, she adds, the problem becomes, “How do you even know something is important enough to tell a person? Anything abnormal should be confirmed” first. Casey now runs pediatric scanning studies and shudders at the thought of “telling parents something is wrong with their child when it isn't.”

Who tells a parent or an adult subject about an incidental finding is an issue, too. Scientists could simply say nothing; there is no requirement in a study scenario. Others might call in a doctor, and if the physician agrees something is suspicious he or she could refer the volunteer to a specialist. Alternatively, a researcher could advise a subject to contact his or her own doctor for follow-up. Or the scientist could call the doctor directly and have him or her address the patient.

Most studies do not include a physician, and participants at the meeting disagreed over whether they should factor a doctor into the cost of a study. “Researchers may be able to detect obvious abnormalities, but they don't have the basic knowledge to make diagnoses,” says Ruth Macklin, a professor of bioethics at the Albert Einstein College of Medicine. Therefore, she maintains, the cost of a doctor or radiologist should be included.

Another vexing issue is that scans used for research are typically not as sophisticated as those in a clinical setting. As a result, the images might be harder to read and interpret by untrained eyes. Buying more expensive machines would further drive up study costs.

By meeting's end the group at least agreed that the consent forms volunteers sign should lay out the possibility that a normal brain might not always look “normal.” The form could ask patients whether they want to know about what seem to be minor findings. And the document should specify that markers of potentially major abnormalities, such as a blood clot, aneurysm or tumor, would trigger immediate attention, whatever that might be. The group agreed to begin drafting guidelines for IFs, including the recommendation to inform subjects and when to refer them to a physician.

Given my own experience, I'm heartened that scientists are paying more attention to this problem. After all, I represent the normal volunteer who was abnormal. I'm glad to know that my dopamine levels are so high that I will probably never get Parkinson's disease. And I've got a picture of my brain, the cyst hogging a good chunk of my cerebellum—the area that controls movement. If I do have any complications from that, it's nothing more than a sore toe on my dancing partner's foot.