Skip to Content

MRI: A Window on the Brain

Advances in brain imaging could lead to improved diagnosis of psychiatric ailments, better drugs, and earlier help for learning disorders.
December 1, 2005

When Bradley Peterson, a psychiatrist and researcher at Columbia University, offered to scan my brain with a magnetic resonance imager the size of a small Airstream trailer, I immediately said yes. I spent 10 minutes filling out a page-long checklist (I lied on the question asking whether I was claustrophobic) and another few minutes emptying my pockets and getting rid of keys, wristwatch, and pen, which could become missiles inside the MRI’s potent magnetic field.

I lay down on a narrow pallet that slid into the machine like a drawer in a morgue. The machine groaned and clanged as it peered inside my skull, then fell silent. With a gentle whir, the pallet slid out, and I relaxed. In about the time it takes to burn a few CDs on my laptop, Peterson was leaning over a screen, showing me a detailed black-and-white image of my brain.

Brain scans like the one I had are now routine, used for everything from detecting signs of stroke to searching out suspected tumors. But researchers like Peterson are pushing MRI technology further than anyone once thought it could go. In the last decade or so, MRI has been retooled to reveal not only the anatomy of the brain but also the way the brain works.

While conventional MRI scans, like the one Peterson gave me, reveal physiological structures, a variation called functional MRI (fMRI) can now also image blood flow over time, allowing researchers to see which areas of the brain are active during certain tasks.

Indeed, fMRI studies over the last few years have provided researchers with startling images of the brain actually at work. A yet newer extension is MRI spectroscopy, another kind of functional imaging that monitors the activity of particular chemicals in the brain – providing different clues to brain function than fMRI does. And most recently, researchers have pioneered an MRI technique called diffusion tensor imaging (DTI) that produces 3-D images of the frail, spidery network of wires that connects one part of the brain to another.

MRI has become, says Robert Desimone, director of the McGovern Institute for Brain Research at MIT, “the most powerful tool for studying the human brain. I liken it to the invention of the telescope for astronomers.” Desimone notes that the arrival of the telescope did not immediately revolutionize the scientific understanding of the universe. That took time, as researchers learned how to use their new tool.

The same thing is happening with MRI, Desimone says. Researchers are just now beginning to realize the potential of these techniques, which were first widely used on humans about 15 years ago. “You’re seeing a lot of excitement in the field,” says Desimone.

Several technical advances have contributed to MRI’s improvement. Topping the list is the development of more-powerful MRI magnets, which enable more-detailed, higher-resolution scans. What megapixels are for a digital camera, teslas, a measure of magnetic-field strength, are for MRIs: the more you have, the better the quality of the image. The newest MRIs generate magnetic fields of about seven teslas, many thousands of times stronger than Earth’s magnetic field and at least twice as strong as those typically used in hospitals. (Some research centers, including the McGovern Institute, have 9.4-tesla MRI scanners for animal studies.)

Another key development is a succession of ever more complex methods of computer analysis. These allow researchers to extract more and better information from scanner data and have improved not just fMRI but also MRI spectroscopy and DTI.

The ultimate aim of brain imaging research is to help explain how the billions of neurons and connections in the brain give rise to thought. But researchers are also applying the new MRI techniques to a more practical, immediate goal: improving the diagnosis and treatment of mental illnesses and learning disorders. The hope is that MRI imaging will provide far more accurate diagnosis of psychiatric diseases whose symptoms can resemble each other, preventing years of suffering for patients put on the wrong medications.

As part of this effort, researchers are using MRI to investigate the causes not only of psychiatric ailments but of all kinds of brain abnormalities and learning disorders, including those often found in children born prematurely. And while attempts to use brain imaging to improve psychiatric health care have met with little success over the last decade, the new MRI technologies – in essence, far stronger telescopes on the mind – are providing fresh hope of finding better ways to intervene.

Bipolar Fingerprint

One of the leaders in the effort to enlist MRI in the diagnosis and treatment of psychiatric ailments is John Port at the Mayo Clinic in Rochester, MN. Port is a neuroradiologist who began his career by studying electrical engineering and computer science at MIT and later earned a PhD in cell biology and an MD from the University of Illinois. So he’s in a good position to research both basic MRI technology and its applications to medicine.

Port’s work on MRI could have broad application in psychiatry, but for now he is concentrating on his particular interest: bipolar disorder. Also called manic-depression, bipolar disorder is characterized by mood swings from wild exuberance to profound depression, with periods of stability in between. X-rays or conventional MRIs show no difference between the brains of people with bipolar disorder and those without it; medical journals are littered with failed attempts to use imaging to find distinctive signs of the disease.

Port thinks a lot of those attempts were scientifically flawed. “I have a list of pet peeves a mile long,” he says. “There are a million studies, but the patients might be on six different medications. So when you see something different, is it the meds? Or is something going on?” Another problem with many earlier studies, he says, is that they included too few patients. “You can’t tell anything from 10 patients. A lot of the research hasn’t been as rigorous as it should be.”

Indeed, despite years of work, neuroscientists still do not know what causes bipolar disorder, or exactly which parts of the brain are involved. That lack of knowledge has severely hampered the search for safer and more effective ways to treat the disease. The principal drugs for bipolar disorder, lithium and Depakote, have been around for decades.

Both were discovered by accident, when researchers trying to do something else noticed that the drugs eased the symptoms of patients with bipolar disorder. And though the drugs can be reasonably effective in some people, doctors have no idea how they work or which patients are most likely to benefit. In order to find better pharmaceuticals, researchers need to be able to target the exact mechanisms or structures involved in bipolar disorder.

Pinpointing the mechanisms could also lead to more accurate evaluation of the disorder. Often, diagnosis in psychiatry is done by a kind of trial and error, in which a psychiatrist makes an educated guess based on the behavior or self-reported symptoms of a patient, prescribes a medication, and sees whether or not it helps. If it doesn’t, the psychiatrist considers a different diagnosis and a different medication, until something begins to work.

“What psychiatrists need is some test that will give them the answer: this patient has the disease or doesn’t,” says Port. He and other researchers hope MRI scanners will offer the definitive diagnosis. And for those in the mental-health profession, that would change everything. “I’m dedicating the rest of my career to coming up with an imaging test that will help psychiatrists diagnose” bipolar disorder and other illnesses, Port says.

Port is one of many researchers now experimenting with MRI spectroscopy, in which software produces an image of the brain based on a spectroscopic scan. The image is made up of individual data points called voxels, cubes analogous to the pixels in a 2-D computer image. Each corresponds to a volume about the size of a kidney bean. For each voxel, Port gets a reading on the presence or absence of certain chemicals that are indicators of brain function.

To understand how MRI spectroscopy works, it’s necessary to understand a bit about how magnetic resonance imaging works more generally. MRI scanners pick up extremely faint electromagnetic signals coming from protons in the atoms of molecules that make up the body’s tissues – in this case, brain tissue.

“Think of it like listening for a pin drop in a thunderstorm,” Port says. Each proton has a magnetic field that points in a certain direction, as the earth’s does. When the MRI is turned on, its magnet aligns the protons’ magnetic fields in the same direction. Bursts of radio frequency energy temporarily knock some of the protons out of alignment. When the protons snap back into place, they release energy, generating a minuscule signal that the MRI’s detectors can pick up. By flipping the protons different ways and measuring various properties of those flips, including the time they take, researchers can identify various tissues and chemicals in the brain.

Using MRI spectroscopy, Port can measure levels of chemicals such as n-acetyl aspartate, which is found only in neurons, or glutamate, which stimulates nerve-cell activity. When Port used the technique across many areas of the brain in bipolar patients and compared the results to those from healthy controls, he came up with a chemical fingerprint that seemed to be an indicator of bipolar disorder.

“When we compared all the bipolar patients in any mood state with their matched normal control subjects, we found that two areas of the brain were significantly different,” Port says. Port and his team also identified changes in many regions of the brains of people with bipolar disorder that indicated whether they were in a manic state or depressed. “We found a chemical measure of the mood state,” he says.

So has Port found the long-sought diagnostic test for bipolar disorder? Does his chemical fingerprint reliably identify people who have bipolar disorder and exclude those who don’t?

Maybe, but he can’t be sure yet. “We think we’re on to something good,” he says, but “we have to check it and make sure it will be clinically useful.” It’s a question of trying the technique with enough patients to be sure that it is statistically valid – that it won’t produce too many false positives or false negatives. It doesn’t have to be perfect, but it has to be good enough to add useful information to what psychiatrists can discern through their traditional methods of diagnosis, interviews, and analyses of patient histories.

If Port is correct, however, and the technique proves itself, it would be a landmark in psychiatric research: a diagnostic test for bipolar disorder. And if the technique works with bipolar disorder, it could be adaptable to other psychiatric illnesses.

Port and others are also experimenting with diffusion tensor imaging. DTI measures water diffusion in the brain. Water flows through the brain as it does anywhere else – along the path of least resistance. In the brain, that’s along the axons, the neurons’ long tails, which convey electrical signals to other neurons. (It’s from the fatty, white insulation that surrounds most axons that “white matter” takes its name; the rest of the neuron, and uninsulated axons, together constitute “gray matter.”)

Port is just beginning to research the technique. But eventually researchers will be able to use “DTI clinically to look for diseases that interfere with white matter – amyotrophic lateral sclerosis [Lou Gehrig’s disease] and schizophrenia,” Port says.

Diagnosing Development

The techniques Port is studying, if they prove successful, will be used in diagnosing people already showing signs of mental illness. But what about others who are predisposed to problems but have not yet begun to exhibit symptoms? Can the MRI technology help to find these people so that they can be helped before symptoms appear?

At Columbia, Peterson is trying to answer that question. He and collaborators are among the first to scan the brains of premature infants – sometimes within days of their birth. The aim is to catalogue the types of brain abnormalities they discover and to devise ways to intervene earlier than ever before to try to correct or compensate for them.

Peterson first became interested in the complications of premature birth about 10 years ago, when he was beginning his psychiatric research at Yale University. He had discovered something very unusual in the brains of people with Tourette’s syndrome. Most of us have asymmetries in our brains – the left side doesn’t exactly match the right. Most of us also have one eye that’s bigger than the other (as portrait photographers will point out) and other minor asymmetries.

But the brains of people with Tourette’s syndrome were different. “In the Tourette’s brain, there seemed to be an absence of asymmetry,” Peterson says. A similar absence of asymmetry had been observed in animals that survived complicated births. Peterson decided to look at children who had been born prematurely. Like Port, he is using the newest MRI technologies to try to obtain information that hasn’t been available before.

There was a reason for his interest. Children born prematurely are at greater risk for learning disabilities and even psychiatric illnesses. Understanding how their brains are different should lead to new ways to help them.

As it happened, Laura Rowe Ment, a pediatric neurologist at Yale, was following a group of 500 premature children born between 1989 and 1992 as part of an ongoing study. Peterson and Ment set up a collaboration. “There were imaging reports suggesting various kinds of problems in the brain – in terms of brain development. But they were uncontrolled, the numbers were small – they were impressionistic,” says Peterson.

Even given their smaller body size, premature kids tend to have disproportionately small heads. “The guess was that brain size would be reduced” later in life, says Peterson. Researchers also speculated that there would be damage to the white matter. Ment’s kids, who were then about eight years old, were especially useful because she and her colleagues had documented everything that had happened to them since they were born.

The first thing Peterson did was use the MRI scanner to determine the size of the eight-year-old children’s brains. The guess was right – their brains were smaller than normal. But the decrease in size occurred only in certain brain regions – the parts of the cortex that govern movement, vision, language, memory, and visual and spatial reasoning. “These regions were dramatically smaller,” Peterson says. The other parts of their brains were normal size, or close to it.

The second guess – about damage to white matter – also proved accurate. There was less white matter in the motor regions of the children’s brains, meaning there were relatively few wiring connections there. And the reduction in volume correlated with IQ scores. “The bigger the abnormality – the more abnormal it was in all these regions – the lower their IQ was,” Peterson says.

The question then was, Did these abnormalities arise at or before birth or sometime later? Peterson started scanning normal and premature infants. The scans of premature newborns showed that they had the same brain abnormalities as the eight-year-olds. “It was so distinctive, the pattern of abnormalities, it’s almost impossible to look at a scan and not be able to tell this is a premature child,” Peterson says.

One of the most salient differences was in the size of the tiny cavities in the brain known as ventricles. “The ventricles are massively dilated, about four times larger in the prematurely born kids than in the term children,” Peterson says. “We saw that in eight-year-olds and in the infants. The tissue around those ventricles is really damaged….It suggests that these babies are having problems in development even before they’re born.” Peterson followed the newborns for two years and then tested them with a kind of IQ test meant for toddlers. The earlier they were born, the more immature their brains were at birth. And the more immature their brains, the lower their intelligence scores.

To neuroscientists, the discovery that premature kids had brain abnormalities made sense. Much of the brain’s growth and development occurs during the last half of pregnancy. Neurons begin life clumped near the center of what will become the brain but soon start to migrate outward. Glial cells, which help neurons communicate, go through a period of explosive growth, accounting for most of the brain’s increase in weight. The neurons extend meandering tentacles, seeking connections with other cells. Billions of connections are made during the last weeks of pregnancy. The axons then develop their coats of white, fatty insulation. By this time, the brain is massively overdeveloped, with far too many wires and connections. So it begins cutting back. It’s as if each connection is tested, to determine its value. The useful circuits are kept; the others are trimmed away, leaving a sleek, efficient machine.

Premature birth likely disrupts these processes – the migration of the nerve cells, the growth of glial cells and white matter, and the trimming. Premature kids have most of the neurons they will carry with them into adult life, but it’s possible they’re not in the right places or properly connected or tested. Researchers, says Peterson, are “intensively testing” these possibilities.

Peterson’s research offers the hope of helping children compensate for whatever brain-related peculiarities they might have. “We want to use imaging to predict who’s going to have particularly difficult problems in the course of development, so we can intervene more effectively,” he says. That intervention might consist of specially designed education programs or physical therapy and other treatments to compensate for physical and emotional difficulties.

When Peterson began this work, his interest was professional. But now he has a personal interest as well. Two years ago, his daughter was born four weeks premature. While she shows no ill effects, he says he watches her, and he worries.

Brainstorming

When Peterson scanned me, he found nothing wrong or worrisome. If I’d had a brain tumor or some prominent abnormality, he would have spotted it. But that’s about all the clinically useful information he could get from a quick scan. If Peterson had put me through the sophisticated scans he uses with the premature infants, perhaps he could have detected some quirk in the way my brain behaves. But because of the large variability in normal brain structure and function, he would not have been able to conclude much specifically about how my brain differs from those of other people.

In the coming years, however, as the technology continues to improve, it may become possible for any of us, with or without obvious illnesses or neurological problems, to learn much more about the state of our brains, our perceptions, and our thinking. “The bad news is that although these techniques are very powerful, they are not where we need to be,” says MIT’s Desimone. “We need to use these MRI magnets in ways they haven’t been used before.”

Desimone’s McGovern Institute has just inaugurated the Martinos Imaging Center. One room at the center houses a state-of-the-art MRI scanner. Beside it is another room that, for the time being, will remain empty. “We have it sitting there for a new device,” Desimone says. He doesn’t yet know what that device will be. “That’s our challenge – to invent it here. The idea is to go beyond where we are now, to the technology of the future.”

Paul Raeburn’s most recent book is Acquainted with the Night, a memoir of raising children with depression and bipolar disorder.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.