New brain scanners promise to deliver images of higher resolution than any now available from a commercial instrument. By using multiple sensors placed close to the head, the device can generate accurate images in less time, which could ultimately aid in the diagnosis of diseases such as Alzheimer's and epilepsy. Medical imaging giant Siemens is developing a commercial version of the technology.

"This might be the biggest-impact development [in brain imaging] for the next few years, especially because Siemens is commercializing it," says John Gabrieli, a neuroscientist at MIT. "If you have a more precise view of the brain, you could take a big step forward."

In magnetic resonance imaging (MRI), a magnetic field generated by a large magnet sends protons in the brain spinning. Specially constructed coils of wire in the machine detect changes in the spin, which differ in different tissue types, as the magnetic field changes. Computer algorithms then use measurements from different parts of the brain to create the anatomical picture.

MRI machines in medical centers typically have up to 12 coils, but the new devices under development have up to 96 coils arrayed in a dense field over the scalp. "A small detector up close is more efficient," says Lawrence Wald, a biophysicist at Boston's Massachusetts General Hospital (MGH), whose team has developed the devices in collaboration with Siemens. "But it only captures a small part of the brain, so you need lots of small detectors spread out over the scalp." Each coil measures a small but highly accurate spin signal from the chunk of brain tissue beneath it. The images are then computationally stitched together to create a high-resolution picture of the brain.

These multichannel devices have already helped some epilepsy patients. In a study using an early prototype, neurologists found abnormalities in about two-thirds of epileptic patients whose previous brain scans had been declared normal, making these patients better candidates for neurosurgery.

Scientists are now using a newer prototype to study Alzheimer's patients. "In diseases like Alzheimer's, where there is not a basic diagnosis based on imaging, we hope that being able to look at smaller alternations in the brain would yield some additional diagnostic information and perhaps allow you to monitor medication," says Wald.

Patients suspected of having Alzheimer's may get an MRI to rule out other neurological causes for their symptoms. But recent studies suggest that subtle neurological changes increase risk for the disease; these changes can include shrinkage of the hippocampus, a crucial memory area, and of parts of the cortex important for memory and higher cognitive function. Detecting these changes requires lengthy scanning sessions to generate high-quality data, making such scans unfeasible in routine clinical practice. "This technology has the potential to change that," says Brad Dickerson, a neurologist at Harvard Medical School, in Boston. While he cautions that routine clinical use is still years off, he says that "we are rapidly moving into a new era where we can use this kind of data to identify abnormalities that are consistent with Alzheimer's."