Functional MRI is a robust research tool for exploring subtle changes in blood flow associated with a variety of activities within the brain. In the clinical arena, however, it is limited for the most part to language lateralization and sensorimotor function brain mapping.
Because the Food and Drug Administration has not approved it for clinical use, fMRI is considered to be a “value added” test. It most commonly is tacked onto a standard anatomical workup of a brain lesion to provide a sort of informal indication of the potential consequences of surgical resection.
But fMRI soon may come into its own. Neuroradiologists are laying the groundwork for obtaining a CPT code for fMRI so they will be able to bill separately for the procedure. Functional MRI researchers are developing better paradigms for studying not only sensorimotor function but memory and mood. And fMRI centers are looking forward to acquiring 3-tesla magnets that will improve their ability to assess cognitive tasks.
The end result will be an active role for fMRI in revealing patterns of brain organization in individuals with neurological and neuropsychological disorders, monitoring the effects of drug treatment for psychiatric illnesses, and uncovering the neurobiology of brain disease.
Current Approaches
There is a strong body of literature showing that fMRI can evaluate language lateralization and map motor as well as visual function. It is not unusual for patients with a tumor near the frontal lobe or the center for control of fine movement to engage in simple and complex exercises within a magnet that has been upgraded with a high performance gradient system to generate a functional preoperative map. Patients may generate words on command, tap their fingers in a prescribed order, or view films or videos.
Researchers nevertheless are still trying to find the best activation paradigms in terms of their content (the actual task given to a patient or subject), the method of their administration (for example, via goggles or earphones that eliminate surrounding noise), and the way the subject is expected to respond to commands.
A series of papers presented at the 1999 meeting of the Radiological Society of North America described some of the latest attempts to refine standard blood oxygenation level dependent (BOLD) fMRI paradigms. These paradigms give patients or subjects an activation task followed by a control task and use statistical correlation methods to analyze the results.
A group at Indiana University has developed a listening paradigm involving a simple narrative text interspersed with the narrative text played backward that selectively activated Wernicke’s area. And a team of investigators from the University of Cologne in Germany used monophasic clicks to visualize activation of the auditory brain stem pathway, which is difficult to examine with 1.5-tesla machines, says Dr. Vincent Matthews, an associate professor of radiology at IU and chief of neuroradiology at IU Medical Center.
“A 1.5-tesla machine is pretty good for looking at cortical function but not subcortical brain function. This group developed a technique for assessing activation in the deep gray matter and brain stem nuclei,” Matthews said.
It is particularly challenging to devise paradigms for studying memory and mood because there may not be much activation in a single patient. That is why some investigators are turning to newer techniques, called event-related fMRI, which collect data obtained over short periods of time in different intervals and add up the results to figure out which part of the brain has been activated during the performance of certain tasks.
Even with the best models, however, neuroradiologists see only small changes in blood oxygen levels with fMRI—on the order of 3% to 5%. The goal is to find deoxyhemoglobin being replaced or overwhelmed by oxyhemoglobin, which comes with an increase in blood flow to an area of the brain. This change in ratio will cause a small increase in signal that cannot be seen with the naked eye.
“We have to study areas of the brain voxel by voxel before and after the activation test and try to pick up these small changes. But at any time in the brain, there can be fluctuations in blood flow due to other normal activities,” said Dr. Pierre Bourgouin, a professor of neuroradiology at McGill University in Montreal.
That’s why Bourgouin sees such promise with the 3-tesla systems that are being purchased by many centers involved with fMRI. The 3-tesla systems will more than double the signal, which should better illuminate the small changes in blood oxygen level and perhaps make it possible to get reproducible data from an individual patient or subject, he said.
The substantially higher sensitivity of the 3-tesla magnets allows a look at performance tasks that don’t produce a big enough signal change to be seen on a 1.5-tesla machine. That in turn shows the lower level changes that occur with cognitive tasks, Matthews said.
“It also allows you to take tasks that are very robust now—motor and visual function—and do them very rapidly so they become more applicable to sick patients who can’t hold still for very long,” he said.
Although Dr. Victor Haughton, a professor of radiology at the University of Wisconsin, believes fMRI is quite effective at 1.5 tesla, he agrees that 3-tesla magnets will facilitate research because they will offer greater precision, anatomic definition, and spatial resolution or contrast resolution, depending on what is desired in an activation experiment.
From Research To Clinical Practice
No matter how sensitive or precise fMRI becomes, it will remain confined to the laboratory until it proves its clinical value. An important first step on the road to clinical practice is to obtain a CPT code for the procedure so the Health Care Financing Administration will pay for functional brain mapping done for Medicare patients. This will take at least two years, Matthews said.
Meanwhile, neuroradiologists are preparing to transfer the discoveries they’ve made in basic research to specific patient populations, such as those with Parkinson’s disease.
Parkinson’s disease is a clinical diagnosis, but the neurological tools used to make the diagnosis are not highly specific. Neurologists in effect exclude other causes of symptoms before reaching the conclusion that a patient has the disease, and the process can take years.
Dr. Michael D. Phillips, an assistant professor of radiology at Indiana University, is applying motor fMRI paradigms to see if they can become reliable as screening tools for the early diagnosis of Parkinson’s disease. Phillips presented preliminary results from an fMRI evaluation of six individuals with early Parkinson’s disease at the RSNA meeting. He compared patterns of motor pathway activation during a complex finger-tapping exam with these patients and eight normal volunteers. The normal subjects had symmetric activation of all brain areas while patients with Parkinson’s disease tended to have relatively symmetrical activation in the precentral gyrus and asymmetric activation in other areas, with decreased motor pathway activation contralateral to the side expressing symptoms of tremor and rigidity. These results suggest that Parkinson’s disease patients may be evaluated with fMRI, Phillips said.
Phillips and Matthews are in the early stages of exploring the concept of functional connectivity, using fMRI to study the background linkages between parts of the brain.
“Even at rest, certain areas of the brain are cycling together. When an area of the brain activates, other areas also get activated or have some input in that function,” Matthews said.
The IU neuroradiologists are far enough along to understand what is happening in some of the normal functional connectivity states, so they and investigators at the Medical College of Wisconsin are beginning to analyze functional connectivity in various patient populations, such as those with dementia.
Functional MRI researchers at the University of Wisconsin have studied a group of patients with Tourette’s syndrome and found differences in functional connectivity—not only in the way areas of the brain are connected but also in the strength of the connections, Haughton said.
Neuropsychiatric Disease
The biggest potential area for fMRI is in assessing brain activation in individuals with psychiatric diseases, Haughton said. “If you an use fMRI to map thoughts, which you can, why can’t you use it to map thought disorders?” he asks.
Already, neuroradiologists around the world are using fMRI to analyze brain function in affective disorders, schizophrenia, autism, and learning problems, such as dyslexia.
Bourgouin has concentrated on mood and the management of mood in normal subjects and patients with major depression. He said that mood is managed predominantly in the lateral frontal cortex and medial prefrontal cortex. In his scientific poster at the RSNA meeting, Bourgouin demonstrated that seven patients with unipolar depression have the same type of activation in the right frontal lobe structures. They also have strong activation at the left medial prefrontal cortex and additional activation in the cingulate gyrus on the right and left sides.
Bourgouin theorizes that major depression may be a form of exaggerated arousal. Perhaps certain parts of the brain are shut down in normal individuals because of negative feedback from other areas.
“Maybe what happens in major depression is that this feedback is lifted so additional areas of the brain are activated. Another theory is that there is an increased general arousal with major depression that calls upon many areas of the brain to manage mood,” he said.
According to Dr. Stephan Pfleiderer of the University of Jena in Germany, patients with unipolar depression may have membrane disorders in the frontal lobe and a loss of mitochondrial DNA. His paper at the RSNA meeting showed a considerable increase in phosphomonoesters, which are precursors of membrane phospholipids, and a decrease in adenosine triphosphate, which may be due to a loss of DNA in the mitochondria, in 14 unipolar depressed patients when compared with eight age-matched controls using 31-phosphorus MR spectroscopy.
If fMRI generates better understanding of neuropsychologic diseases, it likely will move neuroradiologists away from their traditional roles as diagnosticians. What neuroradiologists may offer with fMRI is a means of comprehending the biological processes of psychiatric disease onset and progression.
“You might imagine that a tool revealing the neurobiology of disease may be quite useful in assessing early cases of schizophrenia. There also may be different classifications or patterns of schizophrenia that fMRI may show,” Haughton said.
Functional MRI in addition may be able to mark disease activity and therefore trace the effect of drug treatment for psychiatric illnesses. It might show how the medication affects the organization of the brain, revealing the efficiency of certain drugs, he said.
Bourgouin is taking the first steps toward assessing the effects of drug therapy. His group will perform an fMRI study before and six weeks after treatment for major depression to observe whether the pattern of activation reverts back to normal.
Although the progress of fMRI in the assessment of psychiatric illnesses is difficult to estimate, Haughton predicts that “within a couple of years, we may see important observations with fMRI that suggest the ways in which it can be used to classify or quantitate or monitor neurological and psychiatric disorders.”
If it proves to be possible to draw a map of the neurobiology of neuropsychologic diseases, then fMRI may deliver essential clues to understanding what thus far has been elusive—the natural history of mental illness.
