Mind movies
Paul Thompson has brains. Lots of them. The 33-year-old has degrees in Greek and Latin, mathematics, and neuroscience, and a colleague calls him "the smartest person I know." But we're not just talking about smarts. Thompson really does have lots of brains--about 7,000 at last count.
To see them, step out of the bright Southern California sunshine and into the dark confines of the Reed neurology building at the University of California-Los Angeles, where Thompson has his office and lab. There, in a room behind a heavy glass panel, is a large, humming black computer, and inside are brain images captured by high-tech medical scanners: young brains, old brains, autistic brains, Alzheimer's brains, schizophrenic brains, drug addicts' brains, and a whole bunch of normal ones. "My brain is in there somewhere," Thompson says.
Yet individuality is not what Thompson is interested in. He's mapping brain diseases in large groups of people. By constructing incredibly detailed 3-D images of brains with Alzheimer's and then combining them, he has been able to trace the typical path of the disease and show just how it ravages different parts of the brain over time. Now scientists can see which structures get damaged when--and also see which drugs might keep that damage at bay. In schizophrenia, Thompson's maps have pinpointed a brain region involved in understanding sounds as the first part to be hurt; a common symptom of this mental disorder is hearing voices. His maps are part of a current study of how new and older antipsychotic drugs shield this brain region. "We've never before been able to show these links between brain changes and behavior," says Jay Giedd, a psychiatrist at the National Institute of Mental Health in Bethesda, Md. "So these maps are incredibly powerful."
But they are also still research tools, Giedd cautions, and not useful for actually diagnosing patients. Still, says Andrew Leuchter, vice chair of psychiatry at the UCLA Neuropsychiatric Institute, "one of the most important things about Paul's work is that everything he does helps to trace the earliest signs of disease, because that's the logical place to begin treatment. So much other imaging work is on advanced stages of illness." The key is to see when things first depart from normal. "The brain is the last giant black box of medicine, locked up inside the skull. Paul shows us what it looks like."
Imaging power. Thompson's start in brain science didn't make it seem as if he was going to show anyone anything. "When I first came to UCLA to do my Ph.D., I had trouble in the labs where they were using test tubes and chemicals. A few things got, well, a bit broken." He laughs. "The man running the lab was really quite sweet, but I think I saw a look of relief on his face when my year with him was up."
Things improved when Thompson moved on to neuroimaging. He had always been fascinated by--and good at--advanced mathematics, "and brain imaging is where you can apply math to some tough problems, like understanding what makes people's brains differ from each other. You can collect a lot of brain images, but it's really a mathematical problem to figure out what the patterns of differences are."
Those patterns start out as brain scans taken with a magnetic resonance imaging machine, or MRI. With the MRI, it's as if the brain were an apple sliced into thousands of thin sheets. That's a decent start. But it's hard to get a complete picture of brain anatomy from any one slice. The hundred billion or so neurons that make up our minds form incredibly complex 3-D structures. These clusters of cells vary in thickness, wind around fluid-filled spaces, and are crossed by thousands of creases, canyons, and fissures. A thin cross-section eliminates much of this vital detail and makes it hard to compare whole features in one person's brain with whole features in another.
Thompson's solution was, essentially, to map this complex neurological landscape onto a grid. Like a topographical map of real-world canyons, these grids keep accurate information about the size and depth of brain canyons but gave them a common scale so they could be compared. Using that scale, infinitesimally small features from one brain can be laid on top of similar features from another brain, and size, shape, or thickness can be measured against one another. Do this for a few thousand brains, and you can get an average shape for any brain feature.
The next step is to compare this average brain with an unusual brain, one afflicted with a disease such as Alzheimer's. Because the Alzheimer's brain has lost cells, the two brains won't be the same. It takes a certain amount of bending or flexing to make the two shapes conform; the degree of that bending gives Thompson a measure of the changes wrought by Alzheimer's. He then can blow the brain back up into a lifelike shape, coloring the areas of greatest difference from normal.
Thompson has done this for Alzheimer's brains at first diagnosis, then repeated it every few months. The resulting maps reveal that gray matter loss starts in a region called the hippocampus, a memory area, and quickly moves to the limbic system, which is involved in emotions. Within 18 months, it hits areas in the brain's frontal lobes, which are used to control impulses and make decisions. This nicely tracks the sad sequence of behavior changes that psychiatrists--and family members--have long noted.
"Basically, we're getting a series of snapshots of the brain that show, in fine detail, which structures are getting hit by the disease and to what degree," says Thompson. "And what I think is exciting about this is that it presents us with targets for drugs. We can try a drug, or a combination of drugs, or even diet, and see if it slows down the rate of cell loss in these areas." Indeed, Thompson is now part of a large trial testing Aricept, a commonly used drug that appears to have moderate effects in the early stages of the disease, in combination with the antioxidant vitamin E. The vitamin has been evaluated already with memory tests, showing little effect, but the combination may be more potent. And if it is, that potency will show up in the brain maps.
Real disease. In schizophrenia, the maps show changes "exploding like a lava flow over the brain," Thompson says. The disease often strikes teenagers, causing a frightening mix of hallucinations and psychotic behavior, and it comes on very suddenly. Doctors have puzzled for years over the changes in the brain that might trigger this. A few years ago, Thompson showed that abnormalities in schizophrenics first cropped up in a zone called the parietal lobe. This region integrates input from various senses, like hearing, and passes them on to the rest of the brain. "This really was a revolutionary finding," says Leuchter. "Showing that a brain area involved in auditory processing is abnormal is a lot different than simply saying, 'Oh, they hear voices. It's just in their imagination.' Well, it's not just in the imagination. It makes it into a bona fide illness."
And an illness that can be treated, perhaps better than it has been in the past. Thompson, with scientists at Columbia University and elsewhere, has begun to compare an older drug, haloperidol, with a newer antipsychotic, olanzapine. So far, the newer drug seems to preserve more brain cells. It's also a lot more expensive. But Leuchter and others say that showing it improves brain function helps make the case for using it to penny-pinching health insurers. And that makes Thompson's images not only more than pretty pictures, but true maps to better health.
MAPPING DISEASE
Images reveal how illnesses engulf the brain over months and years.
SCHIZOPHRENIA
The diseased brain first shows damage (pink) at the back, in the parietal lobe. This area helps integrate information from the senses, and damage might cause hallucinations. Over the next five years, the abnormalities move forward.
[Chart labels]
At first diagnosis
2 1/2 years later
5 years later
A BETTER DRUG?
Olanzapine, a newer antipsychotic used to treat schizophrenia, appears to prevent more damage than an older drug, haloperidol, in a preliminary study. Red, green, and yellow areas indicate greater cell loss.
[Chart labels]
Olanzapine
Haloperidol
ALZHEIMER'S
Cell loss, shown in red and white, first hits areas involved in emotions and memory, and then quickly moves to the front of the brain to affect regions involved in self-control.
[Chart labels]
At first diagnosis
6 months later
18 months later
Courtesy Paul Thompson, Kiralee Hayashi, and Arthur Toga; Middle row courtesy Jeffrey Lieberman, Gary Tollefson, Cecil Charles, and Paul Thompson
This story appears in the March 21, 2005 print edition of U.S. News & World Report.