Cover Story
Inside Look
Could brain imaging transform psychiatry?
By Howard Bell
Apostolos Georgopoulos, M.D., Ph.D., dreams of a day when everyone will go to the doctor for a brain check-up. Our brains will be scanned much like our bones are scanned, and those scans will help diagnose psychiatric disorders by visualizing abnormal brain structure and function. Follow-up scans will show whether treatment leads to beneficial changes in the brain. “We have the imaging techniques,” Georgopoulos says, referring to structural magnetic resonance imaging (sMRI), functional magnetic resonance imaging (fMRI), diffusion tensor imaging (DTI), magnetic resonance spectroscopy (MRS), and magnetoencephalography (MEG). “In a few years, we’ll have an image database for what’s healthy and unhealthy.”
Mapping the Brain
The University of Minnesota Center for Magnetic Resonance Research along with Washington University in St. Louis are leading a nine-center consortium in an ambitious effort to learn how the human brain is wired. The Human Connectome Project (HCP) is a five-year National Institutes of Health-funded program that began in the fall of 2010 and will end in the summer of 2015. Similar in scope to the Human Genome Project, the HCP will map the brain’s structural and functional neural pathways and make the data available free to the world’s scientific community.
The plan is to scan the brains of 1,200 healthy adults, including twins and their nontwin siblings, using diffusion tensor imaging and functional magnetic resonance imaging. They will image participants at rest and while performing various tasks. One hundred participants also will be scanned using magnetoencephalography.
The goal is show in unprecedented detail how white matter axons connect different parts of the brain. That information will then be correlated with behavioral testing and genotyping.
The project is predicted to be a leap forward in the emerging field of human connectomics, the measurement of connections between distant regions in the brain, and will set the stage for future studies of abnormal brain circuits in neurological and psychological disorders.—H.B.
|
Human brains have at least 100 billion neurons, and those neurons are capable of an immense number of interactions, according to Georgopoulos, who directs the Brain Sciences Center at the Veteran’s Affairs Medical Center in Minneapolis. And, he says, the information we’re now able to get from very few of those interactions is enough to let us see the difference between healthy brains and unhealthy brains and tell which of the unhealthy brains have schizophrenia, dementia, post-traumatic stress disorder (PTSD), multiple sclerosis, or chronic pain.
Brain imaging could potentially transform many areas of medicine, and it is already being used to help diagnose Alzheimer’s disease (see p. 24). But the implications for psychiatry, where diagnoses have been based primarily on patient interviews, are perhaps greatest, as brain imaging has the potential to render diagnosis and treatment a more measurable science. A number of researchers in Minnesota are currently hard at work trying to realize that potential.
Schizophrenia
Neuroimaging for schizophrenia has been studied more and longer than neuroimaging for any other psychiatric illness, according to Charles Schulz, M.D., head of the University of Minnesota’s department of psychiatry. Thirty years ago, he used computed tomography (CT) to show that adolescents with early-stage schizophrenia had structural abnormalities in their brains.
Other types of imaging have revealed more details. For example, sMRI has shown that people with schizophrenia have a thinner cortical layer, primarily in the frontal and temporal areas that are important to memory, attention, and decision-making. Their hippocampal volume is smaller, too. Functional MRI shows less-efficient neural processing when people with schizophrenia perform memory tasks. And DTI has shown that white-matter fibers are more disorganized in people with schizophrenia than in people without it.
Psychiatrist Kelvin Lim, M.D., a professor of psychiatry at the University of Minnesota, was one of the first to show that white-matter connectivity is abnormal in people with schizophrenia. Lim has found that people with schizophrenia have lower fractional anisotropy values (a measure of white-matter health), particularly in the cingulate, a region of the brain responsible for higher thought processes and emotional control, than people who don’t have the disorder. Other studies have shown decreased fractional anisotropy values in the corpus callosum, which allows the brain’s two hemispheres to communicate and enables sustained attention during complex cognitive tasks. “In schizophrenia, white-matter bundles of axons are not as structurally well-organized as they are in healthy brains,” Lim says. “We believe different brain regions aren’t as well-connected to each other.” No conclusive studies have been done on how or whether medications favorably alter white-matter connectivity.
The Next Step: Treatment
Brain imaging will hopefully lead to numerous new treatments for psychiatric disorders. One of the first to attract attention is deep brain stimulation (DBS), a procedure that was used first to alleviate tremor in Parkinson’s patients and is now being studied for obsessive compulsive disorder, Tourette syndrome, and depression, among other things.
In February, Medtronic received approval from the U.S. Food and Drug Administration for a humanitarian device exemption for its Reclai Deep Brain Stimulation Therapy for chronic, severe obsessive-compulsive disorder. The company is now involved in a clinical trial of the device for depression.
Deep brain stimulation for depression was pioneered by neuroscientist Helen Mayberg, M.D., of Emory University in 2002. Using positron emission tomography (PET), Mayberg identified the region of the brain that appeared to be most involved in depression and then delivered electrical impulses to that area. Some of her patients reported immediate relief.
The University of Minnesota’s Aviva Abosch, M.D., Ph.D., is among the researchers worldwide who are optimistic about the prospects for DBS for depression. “PET imaging shows that a specific part of the subgenual cingulate white matter has increased metabolic activity in patients with major depression, which treatment normalizes,” she says. She also knows DBS needs further study. “We’re not sure how it works.”
|
MR spectroscopy shows abnormally low glutamate levels in many people with schizophrenia, according to Schulz. Glutamate is the brain’s most abundant excitatory neurotransmitter. Low levels may be the reason people with schizophrenia lack motivation, affect, and interest in life in general, even though they are not necessarily depressed. “We also find it interesting,” adds Schulz, “that street drugs like PCP are glutamate-blocking agents that can produce psychotic symptoms that look like schizophrenia.”
Researchers from the university and the VA are using what they’ve learned from imaging to find a biomarker that confirms a diagnosis of schizophrenia earlier so patients can begin treatment soon er rather than later. Starting treatment close to or before the first episode improves long-term outcomes. They also are looking at combining the results of imaging tests with current diagnostic methods. “We already know that combining sMR measures of brain volume with neuropsychological tests is a more accurate way to confirm a diagnosis of first-episode schizophrenia than testing alone,” Schulz says. Including sMR images also helps distinguish between schizophrenia and bipolar disorder, which can be hard to do in the early stages when symptoms are similar, according to Georgopoulos.
Mood Disorders
Brain imaging is revealing fascinating new information about mood disorders, which include depression, bipolar disorder, anxiety disorders, and borderline personality disorder. “We’re a long way from being able to take a picture and diagnose depression,” says Kathryn Cullen, M.D., a University of Minnesota psychiatrist who studies adolescents with major depressive disorder. “But we’re learning a great deal that will one day have practical uses.”
Brain Imaging Techniques
Most brain imaging techniques have been around for decades; but during the last 10 years we’ve seen an explosion of refinements that make these technologies more useful for visualizing the difference between healthy and unhealthy brains.
Structural magnetic resonance imaging (sMRI) – Available since the mid-1980s, this technology uses powerful magnets to produce two- or three-dimensional images of brain structures. It does not show brain activity.
Functional magnetic resonance imaging (fMRI) – Provides a snapshot of brain activity by measuring change in oxygenated blood flow, which increases in regions of the brain where there is increased activity. (Oxygenated and de-oxygenated hemoglobin have different magnetic properties that can be visualized.) fMRI is done while the person is at rest (resting state-fMRI) or doing a mental task (task-fMRI). It has been used in brain mapping since the early 1990s.
Diffusion tensor imaging (DTI) – This type of magnetic resonance imaging measures the health of white matter and identifies disrupted or abnormal white-matter connectivity in different brain regions. The noninvasive technique was developed more than 10 years ago. Whereas fMRI indirectly measures brain activity using a metabolic signal, DTI visualizes white-matter connectivity by measuring the diffusion of magnetically aligned water molecules along axon nerve fibers.
Positron emission tomography (PET) – Measures the location of small radioactively labeled molecules (radioisotope-tagged sugar molecules) in the brain. Areas of higher radioactivity are associated with greater brain activity. In Alzheimer’s disease, PET is used to image decreased metabolism of glucose in areas of the brain affected by the condition. When a molecule that attaches to beta-amyloid protein is used, PET visualizes fibrillar amyloid.
Magnetoencephalography (MEG) – Measures the magnetic fields created when message-carrying sodium and potassium ions speed across synapses between nerve cells. It collects information on brain activity at the same speed as the brain itself operates. Developed in the 1970s to track submarines, the technology is still rare. Only a few MEG machines are available worldwide.—H.B.
|
Using fMRI, she and her colleagues have found reduced connectivity along the fronto-limbic network of neurons connected to the subgenual anterior cingulate in adolescents who are depressed. When intact, this connection prevents excessive emotional reactivity and stress response. She also has used DTI to show impaired connectivity within this circuit. “We believe that disruption of this regulatory circuit underlies adolescent depression and probably adult depression,” Cullen says.
Lim’s DTI studies confirm that the white-matter “hard wiring” between the frontal lobe and limbic regions is not as well-organized in people with depression. “Once we better understand how patterns of white-matter connectivity relate to psychiatric disorders, we can create a diagnostic imaging test for schizophrenia and distinguish between it and depression,” he says.
Whether depression causes abnormalities in the brain or whether the abnormalities cause the depression “is the question that always gets asked,” according to Cullen. Regardless, she believes that brain circuit wiring “went awry” during brain development in adolescents with depression. “Now we need to study the effects medications have on connectivity within this network,” she says. (Thus far, no conclusive research involving imaging the brain before and after medications are given has been done.) “Since adolescent brains are still developing, neuroplasticity may lead to treatments that prevent abnormal neurodevelopment of these circuits,” she explains.
Using fMRI, researchers have found that the brains of people with borderline personality disorder, are characterized by areas of hyperactivity, according to Schulz. “When we show these patients a series of faces expressing a variety of emotions, their limbic areas just light up. We always thought this happened for purely psychological reasons,” Schulz says. “Imaging shows us that parts of these patients’ brains are amazingly over-reactive.”
There isn’t yet a reliable imaging test for bipolar disorder. Measuring the excitatory neurotransmitter glutamate with MR spectroscopy “is currently our best hope for coming up with an objective diagnostic marker for the disease,” says John Port, M.D., Ph.D., a Mayo neuroradiologist, associate professor of radiology, and assistant professor of psychiatry. “So far, we’ve found that people with bipolar disorder have significantly different glutamate levels compared to normal controls. In some areas of the brain, it’s lower and in other areas it’s higher.”
Port is also using MR spectroscopy to measure brain lithium levels in bipolar patients taking lithium, the primary treatment for the disorder. “Half of patients get better with lithium therapy and half don’t,” Port says. “If my new technique pans out, we should be able to tell within a couple days of starting lithium if it will work for a given patient.”
Port hopes that MR spectroscopy will soon help tell whether a patient has bipolar disorder or major depression, which can be difficult to distinguish because depression is a main feature of bipolar disorder. The two conditions are treated with entirely different drugs. “If you put a bipolar patient on antidepressants, they can get worse,” Port says. “So we hope spectroscopy will help us make the right diagnosis so we can prescribe the right medication.”
Glutamate spectroscopy is not yet precise enough to diagnose bipolar disorder in the clinic partly, Port says, because there is considerable similarity in glutamate levels in bipolar brains and in healthy brains. “As scanner performance and our interpretation of results improve,” he says, “we hope to have a powerful diagnostic tool that can be used on individuals. And when we do, it will have significant benefit for public health.”
It Started with Alzheimer’s Disease
Clinicians have been using structural magnetic resonance imaging (sMRI) in the diagnosis of Alzheimer’s disease (AD) since the late 1980s. According to David Knopman, M.D., a Mayo Clinic neurologist and investigator in Mayo’s Alzheimer’s Disease Research Center, when a patient is being evaluated for dementia, sMRI can rule out abnormalities such as tumors, subdural hematomas, and cerebrovascular disease. It also can show localized brain atrophy that is typical of AD. He explains that while a diagnosis of AD by sMRI is not possible with certainty, sMRI can diagnose frontotemporal degeneration (FTD) with confidence.
Positron emission tomography (PET) also can help distinguish between AD and FTD, both of which are characterized by atrophy of the brain. If the radiolabelled sugar used in PET shows decreased activity in the frontotemporal region, the cause of the person’s dementia is FTD. If it shows decreased activity in the posterior temporal, lateral parietal, and medial parietal regions, it is quite likely to be AD. Telling the difference is critical because managing these conditions is quite different, according to Knopman. “For FTD there is no treatment, whereas we have medications such as cholinesterase inhibitors that sometimes slow AD progression but can make FTD worse.”
A breakthrough technique that could revolutionize diagnosis of AD, according to Knopman, is using PET to image fibrillar beta-amyloid protein in the brain. Mayo Clinic’s Val Lowe, M.D., and Clifford Jack, M.D., are using a radioisotope called Pittsburgh compound B that attaches to the beta-amyloid, which can begin to accumulate in the brain as early as 15 to 20 years before the person develops cognitive symptoms of Alzheimer’s.
Brain amyloid accumulation will eventually lead to death of neurons followed by clinical symptoms. Thirty percent of people older than 70 years of age have positive amyloid imaging and, thus, are at greater risk for dementia, according to Knopman. “We’re now trying to determine what percent of those 30 percent go on to develop Alzheimer’s.”
It is believed that people who are not amyloid-positive do not get AD, which makes beta-amyloid a good biomarker for AD, Knopman says.
He and others may one day use amyloid imaging to select patients for AD drug trials. Knopman is currently studying several drugs that would alter or stop the progression of amyloid accumulation at its earliest stages. “Once drugs are developed for treating the mildly impaired or those at risk but with no symptoms, imaging biomarkers like amyloid will be essential. But for now, he says, “we’ve advanced imaging for AD beyond our ability to treat the disease.”—H.B.
|
Post-Traumatic Stress Disorder
Brain imaging using MEG has shown that PTSD changes how the brain works. “We used to debate whether PTSD was real,” says Georgopoulos. “Now we know that it is a brain disease that produces abnormalities in the brain that we can see.”
Georgopoulos and Brian Engdahl, Ph.D., a clinical psychologist and PTSD expert, found that the superior temporal gyrus of the right hemisphere, which is involved in causing us to relive past experiences, interacts with other parts of the brain very differently in people with PTSD than in healthy people.
Pinpointing the area of the brain that is hyperactive in persons with PTSD is one step toward finding a diagnostic biomarker for it, according to Jose Pardo, M.D., Ph.D., director of the VA’s Cognitive Neuroimaging Unit. “Once you have your biomarker, you can test the effectiveness of treatments,” he says.
Engdahl and Georgopoulos are in the early stages of using MEG to study the brains of people with Gulf War syndrome, which has a variety of unexplained physical and psychological symptoms, and of those who have suffered mild traumatic brain injury (TBI). “You can see the difference between someone who has TBI and someone who doesn’t,” Georgopoulos says. “We’ve even identified a subgroup of veterans who were pronounced cured of TBI whose brains are not normal whatsoever.”
As for Gulf War syndrome, Georgopoulos says, “we’re going to find out the neural basis of the symptoms using MEG. If we can image it, then we can test the effectiveness of therapy by imaging before and after treatment.”
Drug Addiction
Brain imaging could have profound implications for those working with people with drug addictions or babies born with fetal alcohol syndrome. Using DTI, Jeff Wozniak, Ph.D., a pediatric neuropsychologist at the university, has found that adolescents exposed to alcohol in utero have significantly disrupted white-matter connectivity, especially in the corpus callosum. Using MEG, Georgopoulos has found that the brain activity of chronic alcoholics undergoes rapid changes after the person goes through detox. “Within seven days of sobriety, the brain activity flipped with a dramatic shift toward normal,” he says. Likewise, using DTI, Lim has found decreased connectivity in frontal areas and in the corpus callosum of cocaine addicts, a finding he thinks helps explain their impulsivity. “We know cocaine constricts blood vessels,” Lim says. “We believe reduced blood supply alters white matter and reduces connectivity.”
Still a Basic Science
Whether these imaging techniques make their way into psychiatrists’ offices any time soon remains to be seen. For now, says Georgopoulos, all the work being done is still at that basic science level, providing researchers greater understanding of the areas involved in specific psychiatric disorders. “One thing we have learned,” Pardo says, “is that terms like depression, schizophrenia, and dementia are gross descriptions that include many subtypes.”
In order to better understand those subtypes, imaging technology and the methods used to analyze the images they produce need to become even better than they are now. “As they do,” Lim says, “we’ll more precisely locate microstructural abnormalities associated with specific psychiatric disorders. For now, there’s considerable overlap in findings that makes differential diagnosis difficult.”
Even if neuroimaging does improve to the point where it is ready to be used in the clinic, cost and access become factors. A PET scan, for example, costs around $3,000. And the VA’s MEG imager, which cost $3 million, is the only one in Minnesota with 248 sensors.
“It’s not practical to scan millions of people for depression,” Cullen says, “especially since we’re already pretty good at diagnosing it.” Instead, she says, scanning might be used selectively for questionable cases where symptoms overlap different diagnoses. And sometimes it will be useful for prescribing the right medication the first time, which can be difficult now, according to Cullen, because “each of the antidepressants we have works for only 60 percent of patients.”
Cullen is trying to develop imaging “neuroprofiles,” subtypes of depression based on the areas of the brain that are affected correlated with results from traditional clinic evaluations. The profiles would guide treatment choices by identifying a more precise treatment target in the brain. “That way, we get patients on the right medication for them without wasting time on several drug trials,” she says.
But translating such basic research findings into clinical tools requires more and bigger studies, and getting grants to perform them is a challenge, according to Pardo. “The grant people ask us, why image if you can’t treat it? We tell them we can’t develop treatments if we don’t image. Mental diseases are not what people like to fund.”
Clearly, more work needs to be done before we will be able to use brain imaging to diagnose and treat psychiatric disorders. But with each study, we are moving closer toward that capacity. “Right now, everybody is investigating their own small piece with their favorite technique,” Georgopolous says. “Each gives additional information and each has its advantages and disadvantages. But psychiatry is like the rest of medicine. You don’t look at just one thing. You look at everything you’ve got.” MM
Howard Bell is a medical writer and frequent contributor to Minnesota Medicine.