"You can't replace the brain. But that's our job -- to learn to do that."

The paradoxical statement comes from Dr. Richard Jed Wyatt of the National Institute of Mental Health, and he added one word: "Someday."

The feat is still impossible, because the surgeon would have to make severed nerve cells grow back together. They do not grow back together in nature. A brain transplant -- really a body transplant, because the brain would contain the personality of its donor, not its recipient -- would involve grafting the nerve cells of the brain onto a new spinal cord.

But Wyatt is among a group of scientists and surgeons who have been making remarkable progress toward a part of that distant goal by transplanting some important brain cells. In the past few years, they have implanted bits of new tissue into scores of animal brains to try to correct some of the disorders that afflict human brains.

Last spring a Swedish neurosurgeon, Dr. Erik Olof Backlund, made such a transplant into a patient with Parkinson's disease, the "shaking palsy," in an effort to relieve the symptoms of the disabling disease. The jury is still out on his attempt, but there has been enough success in animals to make some researchers think some human success could be only five to 10 years away.

Wyatt's contribution to that success is carried out in a laboratory in an old red brick building at St. Elizabeths Hospital here, with the help of white rats.

Wyatt, Dr. William Freed and associates first destroy a small part of such a rat's brain, one of two darkish areas called the substantia nigra. This deprives the rat of dopamine, a vital chemical produced there that regulates movement. Without dopamine, the rat walks or staggers in circles rather than a straight path. It has been given an approximation of the human Parkinsonism that occurs when there is a dopamine shortage.

The animal is given time to heal. Then Wyatt takes either a bit of substantia nigra from the brain of a fetal rat or a part of the adrenal gland (just above the kidney) from another adult rat.

The material is implanted in the crippled rat's brain. If it is substantia nigra, its fibers begin growing and it is soon part of the brain, making dopamine. If it is transplanted adrenal, it does not grow into the brain, but it survives and releases dopamine. In either instance, the crippled rat will now walk more normally in most cases.

Scientists at the universities of California, Colorado and Rochester, at Yale, Purdue and Mount Sinai Medical Center in New York, at the Medical University of South Carolina and at Sweden's Karolinska Institute and University of Lund are doing similar experiments. Some transplant brain cells that make hormones -- either an important sex hormone called LHRH or vasopressin, which maintains the body's water balance.

Wyatt and Freed's group are also trying to make blind rats see, by transplanting fetal eyes into the optic pathway within the brain. The eye and brain are the only organs that do not quickly reject foreign tissues.

"We think the really major contribution of all this will be understanding--how the brain connects with itself, how it grows, how one cell finds its target," Wyatt says. "What is the brain's plasticity, its capacity for physical change? The brain is not cast in concrete--there is constant physical change. But we don't know how much or, when parts are damaged, how much they can regenerate and how the brain then can be aided."

The experiments being carried out by Wyatt and his colleagues around the world carry huge implications for the treatment of hormone deficiencies and some disorders that affect the body's ability to move. It might one day be possible, for example, to restore parts of the brain or nerve tracts damaged by accidents or strokes.

Scores of scientists are working on the problem of nerve cell regeneration as well, hoping one day to prevent or allay paralysis in some of the thousands of persons whose spinal cords are severed in accidents.

"The neurosurgeons are interested in what we are doing," Wyatt says. "They keep asking, 'Have you got anything we can use yet?' " The answer so far is "Not yet."

But someday.

The first "surgeons" to approach the brain were Cro-Magnons.

Starting in the Upper Paleolithic age (from 60,000 to 10,000 B.C.), ancient healers in Europe, Egypt and South America treated injuries and various mysterious afflictions--probably epilepsy, paralysis and mental illness--by opening the skull to release "bad spirits." With sharpened flints and, later, bronze knives, they cut and scraped through the skull to chisel out circles or squares of skull bone, certainly reaching the brain's outer layer.

Many of the patients survived the crude surgery. Archeologists have excavated skulls that bear both the marks of the crude incisions and new bone growth that filled them in later years.

And chances are that some patients improved. Just opening the dura, the brain's outer covering, may release excess blood or fluid in some cases and relieve pressure on sensitive brain cells.

But neurosurgery and neurology (the non-surgical, medical care of the brain and nervous system) are still often called "the gloomy sciences." Despite more than 60 millennia of brain surgery, many patients still continue to suffer or die.

But the picture is changing dramatically--especially in the last decade.

The early 1970s brought CAT-scanning: detailed computer reconstructions of series of X-ray slices of the head and brain, taken from many angles to pinpoint problems.

Among those who marvel at the speed of the changes is Dr. Bruce Ammerman, a Washington neurosurgeon. "Say you confront an epidural hematoma a blood clot on the brain's outer layer, the dura ," he says. "Ten years ago you had to do a cerebral arteriogram a dye injection followed by old-style X-rays . It took an hour or two. The patient often deteriorated. Now you get a diagnosis within 30 minutes. The accuracy and speed in getting the patient to the operating table can be lifesaving."

Dr. Donlin Long, chief neurosurgeon at Johns Hopkins University Hospital, is looking forward to the day when neurosurgeons can count on even more sophisticated diagnostic tools, among them the so-called PET scanner that is a leap ahead of the CAT scanner.

"It will show us not just anatomy but functioning," said Long. "In both epilepsy and brain tumors, we should be able to define normal and abnormal areas much more sharply."

In removing brain tumors -- which are frequently cancerous and frequently fatal -- the ability to discern precisely the difference between healthy and unhealthy tissue can be critical.

"One problem in operating on them now," Long says, "is that you don't really know where to stop, so you may remove some healthy brain cells." Removing healthy brain cells can mean saving a person's life at the cost of paralysis, blindness or other abnormalities.

Johns Hopkins surgeons are already injecting the brains of some tumor patients with a fluorescent dye that stains tumor cells but not normal cells. The surgeon then wears a special headlamp under which the tumor shines a bright, telltale green.

Ultra-sound is giving University of Chicago surgeons another kind of window to the brain. They are trying sound waves to find tumors, cysts, abscesses, aneurysms -- swollen or malformed blood vessels -- and fragments imbedded by accidents.

Thin catheters may be threaded into the brain through its blood vessels to reach some aneurysms and seal them off. New anesthetics give surgeons more time for such elegant techniques.

Another advance is stereotaxis -- or guided positioning -- rather than merely searching around in the brain to seek an abnormality. A CAT-scan of a tumor can now be followed by mounting a metal guide on the patient's head, with a hollow steel rod leading straight to the tumor site. The rod guides the surgeon and gives him an enlarged, microscopic view as he wields fine instruments through the same pathway.

The microscope may be the most useful "new" instrument. Microscopes have been used in some operations, notably eye and ear surgery, for years. But only recently have brain surgeons been given truly useful operating instruments to expand their vision by up to 40 times. With new micro-instruments, miniature scalpels and forceps on the ends of long probes, neurosurgeons can now repair blood vessels no wider than a pencil lead with sutures even thinner than a hair.

The bottom line of all these advances is that more patients are surviving even the most delicate brain surgery.

At George Washington University Hospital one recent day, Dr. Arthur Kobrine embarked on a spectacular example of the new neurosurgery. The procedure is called revascularization -- what it means is that the surgeon will create a new blood supply line for a brain starved of oxygen, to ease the symptoms of a stroke or try to prevent one, or to repair a dangerous aneurysm and prevent a cerebral hemorrhage.

The patient was in his early 60s, suffering from a total blockage of the carotid artery on the left side of his neck. The blockage had cut off blood flow to the left side of the brain, causing periodic small strokes, with paralysis and loss of speech.

Kobrine opened the patient's temple to find the slender left temporal artery, the artery you can feel pulsing at your temple. He freed one end of it and applied micro-clips to staunch bleeding. This vessel, three one-hundredths of an inch in diameter, would be the graft.

The surgeon opened the scalp and skull to make a silver-dollar-sized opening or craniotomy. He peeled away dura to find the pinkish-gray brain underneath.

Kobrine quickly identified the artery that was not getting enough blood because of the blockage. In fine maneuvers, he clipped off a section of this spaghetti-like vessel, then sutured one end of the fine temporal artery into its side.

The walls of the blocked artery were as thin as paper and virtually transparent. The diameter of the suture material the surgeon was using was 22 microns -- eight millionths of an inch. All the suturing and knotting had to be done through the microscope, at 25 to 40 power.

"This is the most delicate, tedious operation we do," Kobrine said. "If you pull too hard, the suture breaks or it goes through the vessel wall."

Neither happened, and finally the temporal artery was attached to its host vessel. The clips were removed, and the artery from the temple quickly filled the pale and starved cerebral artery with fresh blood. Both pulsated.

"Now we're simply bypassing the old blocked area," Kobrine said. "Now we're bringing the brain new blood. That's the bottom line."

"The future, the next 10 to 20 years," says Dr. Donlin Long, "is going to bring a great increase in what we call 'functional surgery.' " To Long, the phrase carries a golden promise that surgeons, using their ever more sophisticated tools, will be able to alter the brain's function -- to halt chronic pain, for instance.

But the idea of surgical intervention to change brain function frightens many persons.

The fear stems from the decades of the 1930s and '40s -- and even later -- when thousands of operations attempted to treat mental illness by cutting brain fibers. This "psychosurgery" often blunted memory and personality as well. The episode was a well-intentioned but sorry one in psychiatric history.

This kind of surgery is largely confined now to a few hundred far more refined operations each year to suppress either intractable pain or some cases of violent, often self-mutilating mental illness treatable in no other way.

What Long foresees are very different techniques, in which neurosurgeons will implant electrical stimulators in the brain or nerve tracts to control pain or to manage epileptic seizures or grotesque movement disorders.

Much of the effort to control movement problems has been led by Dr. Irving Cooper of Westchester County Medical Center in New York. Johns Hopkins' Dr. Long, with Robert Fischell at Johns Hopkins Applied Physics Laboratory, has developed an implantable and rechargable "neuropacemaker" that a patient can regulate to relieve pain or involuntary movement.

Long believes the next step will be to use the brain's own neurotransmitters--the message-carrying chemicals of the brain -- to do the same jobs. "I think we'll be able to implant pumps or instruments to deliver these chemicals to the brain or spinal cord," he said.

The kinds of sophisticated electrical or chemical devices that excite the neurosurgeons, however, raise questions in the minds of others, suggesting a more fearful kind of "brain control." In the wrong hands, it might be just that.

More than 10 years ago Dr. Jose Delgado of Yale implanted electrodes in brains of animals to train them. By remote radio control, he was able to stop an angry bull's charge and make an excitable chimpanzee less aggressive.

By putting electrodes in a human patient's brain, he could force the patient to close his hand into a fist every time the current stimulated a certain area. Other patients, similarly stimulated, suddenly "heard" a noise or suddenly turned their heads. They were entirely convinced their actions were spontaneous, not stimulated.

Delgado spoke of using his electrodes to treat such common illnesses as anxiety and obsessive behavior. But in a powerful book called "The Brain Changers," writer Maya Pines objects that such talk "makes all of us possible candidates for brain stimulation." Tie some implanted electrodes to a computer that stimulates some specified behavior, she says, and you have "all the elements of a fiendishly efficient system, the most effective form of conditioning known."

The potential for abuse is one of the questions that must be weighed as science slices into the mind.