It was a race against time at the National Institutes of Health research hospital as a medical team hurriedly placed Barbara Bath in the padded metal bed of a space-age machine that was about to invade her brain.

The rush was not a matter of medical emergency but of nuclear necessity, dictated by rapidly decaying isotopes sending radioactive messages from inside her head. Bath was about to have her mind read by a doughnut-shaped machine that, with the help of a sophisticated computer, would produce a picture for the hovering scientists.

The painless procedure posed a minimum of risk to Bath, 44, a graduate student and volunteer in a National Institute of Mental Health (NIMH) experiment involving the hottest new technology in neuroscience research--Positron Emission Tomography, or PET for short.

The excitement about the potential for PET follows the revolution ushered in a decade ago by Computerized Axial Tomography (CAT), a new X-ray technique for penetrating the skull to visualize the structure of the brain tissue inside. But while CAT shows structure, PET shows function.

In the past, limited information about brain function has come indirectly from chemical markers in the urine, blood or spinal fluid. Studies of accident victims and autopsies of brains have provided other insights, but neither offers a look at living, healthy tissue.

The PET, however, can take a kind of "radiological biopsy" of the biology of various parts of the brain, says physicist Rodney Brooks of the National Institute of Neurological and Communicative Disorders and Stroke.

It does so by tracing radioactive isotopes bound to biologically active molecules -- such as sugar, oxygen, carbon or nitrogen. For example, the isotopes attached to molecules of glucose, or sugar, will pinpoint the area of highest brain activity, under the principle that the brain's busiest areas need the most energy.

Working with mice, rats and monkeys, NIMH researcher Louis Socoloff measured this by tricking the animals' brains with a sugar "decoy." He found that when he radioactively tagged a chemical relative of sugar -- one that the cells would take up but not break down completely -- the chemical left behind a record of its destination.

Researchers at the University of Pennsylvania and New York's Brookhaven National Laboratory adapted this approach for use in man. A team at Washington University in St. Louis pioneered the development of the PET instrument itself.

At the NIH hospital, Bath sat in an acoustically sealed, darkened room with her eyes closed, receiving a series of shocks in her right arm. She was part of a healthy control group in a comparative study of schizophrenic patients, following up suggestions that schizophrenics have diminished pain response in addition to severe psychosis.

On her head, a mass of 16 EEG wires also measured electrical activity responses on the outside surface of her brain.

After receiving the shocks, Bath was taken to the PET room where she lay for more than an hour as the machine, starting near the top of her head and moving slowly downward, took activity readings from seven horizontal "slices" of brain tissue.

The computer converts the information into colored or black-and-white pictures in which "hot" areas indicate the most intense activity and "cold" areas the least.

Bath's brain showed greater activity in the left cortex, the area thought to be in charge of interpreting sensory messages. But it also showed intense activity in the right cortex, which has been linked with mood, apparently showing an emotional response to the shock as well.

The brains of schizophrenics tested thus far showed less activity in both areas. "This is new and exciting," beamed NIMH psychiatrist Dr. Monte Buchsbaum. "It links the pain system to theories about mood."

Just before Bath, the PET bed was occupied by another experimental subject, a spry 72-year-old Virginia man participating in a National Institute on Aging study comparing healthy versus senile brains.

And in the new wing where a more advanced machine called Neuro-PET is housed, two patients with "gliomas," or brain tumors, were wheeled in from the hospital ward for a neurology institute project. The new instrument is proving particularly useful in determining which cancers are growing most quickly. The more "aggressive" the tumor, the more sugar it consumes for energy, explained Dr. Nicholas Patronas.

While researchers emphasize their work is still experimental, studies at NIH and around the country are already providing some patient treatment payoffs.

At the University of California at Los Angeles, PET pioneer Dr. David Kuhl and his colleagues are providing a more specialized diagnosis of epilepsy, using both the EEG and the PET to detect small zones in the patients' brains that are responsible for their seizures. On the PET pictures, these areas seem low in activity before a seizure and overactive during one.

The visual map helps them decide whether to do surgery on patients who are no longer helped by medication. In 25 cases in which surgery was performed, Kuhl said that 20 showed an "excellent outcome with the disappearance of seizures" and no apparent adverse effects.

There is also the suggestion, says Kuhl, that PET has diagnostic potential for Huntington's Disease, a hereditary affliction that strikes in the late 30s, with a subsequent deterioration of mind and body.

In patients with early symptoms, the PET picks up signs of abnormal activity in a portion of the brain thought to be associated with the disease. The PET may find similiar activity in patients at risk of developing the disease before any symptoms have shown up. It will take long-term followup to see how accurate the signals are.

In the future, doctors also plan to use PET to develop better drug treatments for patients with brain disorders, tagging medications with radioactive tracers to discover where in the brain the drug is most active.

Because the process is expensive, time-consuming and limited by the need for a cyclotron to generate radio-isotopes, PET is available now to only a select group of scientists around the world as a research tool. A single PET scan can cost upward of $2,000.

PET researchers at UCLA, Johns Hopkins University and other institutions hope to spin off a less costly technique called "single photon emission computed tomography" or SPECT, which uses commercially available radio-isotopes and could be installed in facilities that have no access to a cyclotron.

Instead of measuring energy metabolism, SPECT maps blood flow, which also goes up in areas of the brain that are active, explains Kuhl.

Improving the PET itself, scientists have also sought to overcome the drawbacks of the commonly used radioactive sugar approach, which provides only a slow-exposure picture of average brain activity over a half-hour period. They would like to have something more on the order of a series of snapshots, showing the moment-to-moment activity that accompanies individual thoughts.

With on-site cyclotrons (NIH expects to have one installed by 1984), PET researchers are already taking advantage of faster-acting radio-isotopes which, combined with the latest machines, can produce pictures of a minute or less of brain activity.

"Imagine how lightning-fast your mind must change from one type of mental activity to another," said Dr. Herbert Meltzer of the University of Chicago School of Medicine. With the faster machinery, he said, "we have a better chance to see a signal from disturbed mental functioning or in response to a stimulus."

"Ideally," he said, "we would eventually like to do a moving picture."