Researchers at the Mayo Clinic in Rochester, Minn., have developed a
brain-imaging technique that could prove to be a potential diagnostic tool for
bipolar disorder. Preliminary data, presented last month at the annual meeting
of the Radiological Society of North America in Chicago, identified
differences in the brains of patients with bipolar disorder, compared with
those of control subjects.
The research team, led by John Port, M.D., Ph.D., an assistant professor of
radiology at Mayo, used magnetic resonance spectroscopic imaging (MRSI), a
form of magnetic resonance imaging that allows clinicians to image not only
anatomical structures, but also physiologic functions.
Funded by the Research and Education Foundation, Port has submitted the
work to a peer-reviewed radiology journal.
"Anatomical imaging has truly failed to help out psychiatry,"
Port told Psychiatric News. "Structurally, the brains of
psychiatric patients and `normal' patients are largely
indistinguishable—you simply cannot reliably image the
differences." Port noted that several studies have found small
differences, but there has been no consistent trend that has pointed to a
future diagnostic tool.
Previous MRSI studies of patients with mental illness have been plagued by
two "fatal flaws," Port said, which made the search for reliable
findings analogous to looking for the proverbial needle in a haystack. First,
earlier studies used metabolite ratios comparing two different physiological
byproducts in the brain. However, Port pointed out, if the ratio is different
in a patient with mental illness, what exactly is different—did the top
half of the ratio change, or did the bottom? Using ratios simply doesn't
answer that question.
Second, most earlier studies were limited to images of a relatively small
amount of brain tissue.
"In previous studies they would look at areas [within the brain] that
were thought to be involved in bipolar disorder, for example, yet what if the
abnormality they were looking for actually occurs somewhere
else—somewhere they didn't look? The studies were subject to sampling
error," Port said.
Port and his team set out to avoid these "fatal flaws." His
group measured actual brain metabolite concentrations, not just ratios of
metabolites, and they did so using a technique developed by researchers at
King's College in London that allows the collection of hundreds of voxels at a
time, rather than the one or two voxels commonly imaged in previous studies.
(A voxel is a unit of volume of tissue imaged during a scan.) In addition, the
Mayo group used a "three tesla machine," a magnetic resonance
scanner with the strongest magnet currently available and nearly twice as
strong as that used in previous studies.
Linking form and function, magnetic resonance spectroscopic imaging
(MRSI) allows simultaneous imaging of brain anatomy and physiology. Images
courtesy of John Port, M.D., and Radiological Society of North America
"That allowed us to get the information we needed in about half the
time," Port explained, adding that minimizing the time that patients
with bipolar disorder were required to spend in the scanner was a "real
"The fourth piece that makes our study really unique is that we
recruited only drugnaïve
patients," Port added, "many of the previous studies were done
with patients on a wide array of medications, which of course is a serious
potential confounding variable."
Port and his team scanned the brains of 21 adults with bipolar disorder who
had either never been on medication (having been newly diagnosed, for example)
or had not taken any psychiatric medication for at least the eight weeks prior
to the scans. They also scanned the brains of 21 volunteer adults who had no
psychiatric history. The control subjects were matched by age, sex, and
handedness. Study participants ranged in age from 18 to 54; there were 26
women and 16 men.
The researchers scanned between 60 and 70 regions of the brain at a time,
gathering thousands of data points. That allowed them to analyze 14 areas of
the brain for differences in four major brain metabolites: myoinositol,
creatine, choline, and N-acetylaspartate (NAA).
Specifically, myoinositol concentrations were significantly higher in the
right frontal white matter of bipolar patients, compared with control
subjects. In addition, both creatine and NAA were significantly lower in the
right lentiform of bipolar patients, compared with normal control
"But after the initial differences were seen, we wanted to know
whether the subtype of the disorder—bipolar I, II, or NOS—made any
difference," Port explained. "Would we be able to distinguish
The team's analysis revealed that bipolar I patients had significant
differences in brain metabolites, compared with bipolar II or NOS
"What that tells us is that bipolar I seems to indeed be a different
disease than bipolar II or NOS," Port said. "We looked at left
caudate choline levels, right parietal white-matter choline, and NAA levels,
and the bipolar I patients had the largest differences in metabolite
concentrations compared with the control subjects. It makes sense—it
fits with the clinical picture. Bipolar I is clinically the more severe form
of the disorder."
Finally, Port said, the team looked for differences in metabolite
concentrations tied to the patients' mood state when they were scanned.
"The most significant differences were seen between bipolar patients
in a mood state—manic or depressed—and those who were
euthymic," Port said. "Again we found different concentrations of
different metabolites, which reinforces that these abnormalities correspond to
mood states—a fingerprint for mood state, if you will."
Intriguingly, he noted, they found little difference between patients who
were depressed and those who were manic, but both polar states were different
from those in subjects who were in a normal mood state.
The team is continuing to refine the technique, and Port plans to submit a
grant proposal to the National Institute of Mental Health for funding to study
the technique in a larger, well-controlled population.
"What I'd really like to look at is patients who have bipolar
disorder and compare those with both a normal, healthy control population and
a group of patients with major depression," Port said. "That way,
we could see if we can tease apart bipolar from major depression."
"Most diseases you can diagnose with a blood test, a scan, something
tangible," he noted. "That has been elusive in psychiatry.
Hopefully this represents progress on that front."
More information on Port's work is posted online at<www.rsna.org/rsna/media/pr2004/pr_bipolar.html>.▪