Each detector is positioned over a specific part of the brain, generating a high-resolution image of that tiny piece. (With conventional scanners, a single detector picks up signals from the entire brain, generating a much lower signal-to-noise ratio.) The sections are then put together to form a detailed picture of the whole brain. "You can make individual channels smaller, so they have greater sensitivity to parts of the head that are close to that channel," says Graham Wiggins, a research fellow in radiology at Harvard Medical School, who helps to engineer the arrays.

These array detectors are fast becoming standard practice -- Siemens now offers a commercial 32-channel array system. But Wiggins' team, in collaboration with Siemens, is engineering much denser arrays with even better resolution, that they say would be useful for understanding a number of neurological diseases. "We might be able to see white matter anomalies connected to multiple sclerosis or early changes related to Alzheimer's disease," he says.

Cleveland's Najm says the higher resolution is key to understanding epilepsy and other diseases. "I think we need to continue to improve imaging ability, until we get to a cellular resolution about 10 to 15 microns, where a cell will be seen as cell. Then we can really see the cellular organization." He adds that combining MRI with other types of imaging, such as molecular imaging, which allows scientists to examine activity inside a cell, will also help identify abnormalities associated with epilepsy.

Grant and colleagues are now trying to couple these high-resolution structural images with other technologies that measure the electrical activity of a seizure, to determine exactly how structural glitches generate seizures.

Extremely high-resolution imaging has its downside, too, though. It's possible, for example, that many people have tiny cortical flaws that don't noticeably affect their health. "The more sensitive the imaging studies get, the more you have to think about how findings are related to the illness," says William H. Theodore, chief of the Clinical Epilepsy Section at the National Institutes of Health in Bethesda, MD. He adds that the value of new imaging technologies also depends heavily on the person reading the scan. "An expert reader of a somewhat inferior MRI scan may do just as well as a less expert [reader] of a better MRI scan," he says.