New magnetic imaging technique can show brain activity

MURRAY HILL, N.J. -- A new technique for imaging the human brain that yields images of brain function, rather than anatomy, has been demonstrated by a team of scientists from AT&T Bell Laboratories and from the University of Minnesota Medical School.

The technique, a variation of Magnetic Resonance Imaging (MRI), is detailed in a paper published in the July 1 issue of the Proceedings of the National Academy of Sciences. Instead of MRI's standard images of different types of brain tissue, the AT&T/Minnesota images show the precise locations of increased neuronal activity within a living brain.

"Our method adds an entirely new dimension to MRI images of the brain, allowing us not just to see the brain, but to see it working," says Bell Laboratories biophysicist Seiji Ogawa, who pioneered the development of the new method. "By imaging the brain in this way, we can map the specific sites of certain types of mental activity."

"We've known for many years that our brains contain discrete anatomical areas that are specialized for performing different types of thinking," said David Tank, head of the Biological Computation Research Department at Bell Laboratories in Murray Hill, N.J.

"Our new method can be used to map these areas in a normal, awake person with an accuracy that was impossible to achieve with previous brain imaging methods.

"We have already demonstrated that we can pinpoint--down to the scale of millimeters--the precise parts of the brain that are activated by a simple visual stimulus and during coordinated hand movements," says Tank.

"We believe the method can also be used to image brain areas involved in higher cognitive function, such as multiplying two numbers together or creating mental images."

Tank's department is attempting to learn from biological neural networks how to design better computers.

"Medical researchers and clinicians are excited about using this method to map human brain activity," according to Kamil Ugurbil, director of the Center for Magnetic Resonance Research at the University of Minnesota.

"It will be possible to provide more accurate images of the brain for neurosurgeons to use in planning surgery, and possibly to localize those brain areas that are active during epileptic seizures.

"What is really exciting about the new technique is that it is completely non-invasive: it requires no surgery and poses no health risk to the subject," Ugurbil added.

Called Blood Oxygen Level Dependent (BOLD) imaging, the new method detects increases in blood flow to active areas of the brain. Specifically, BOLD imaging detects changes in the amount of oxygen that is bound to the hemoglobin molecules in each area of the brain.

"It is now well recognized that MRI provides exquisite images of brain anatomy and pathology," said Ugurbil. "This new development of functional mapping brings a totally new and novel dimension to the use of MRI. No other method has proven to be so versatile, comprehensive and powerful in biomedical research and clinical medicine.

In addition to Ogawa, Tank, and Ugurbil, the other authors of the PNAS paper are Ravi Menon, Jutta Ellermann, Seong-Gi Kim, and Hellmut Merkle, all research scientists at the Center for Magnetic Resonance Research.

Functional human brain imaging using the BOLD contrast method has also recently been demonstrated at lower magnetic field strengths by researchers at Massachusetts General Hospital and the Medical College of Wisconsin.


"As the body increases blood flow to the active regions of the brain, that increased blood flow is detected by looking for pockets of high-oxygen blood," explains Seiji Ogawa. "That's where Magnetic Resonance Imaging comes in. Hemoglobin without oxygen--known as deoxyhemoglobin--is paramagnetic, meaning that when it is placed in a magnetic field, it slightly increases the strength of the field in its vicinity.

On the other hand, hemoglobin containing oxygen is not paramagnetic, so when a lot of deoxyhemoglobin is present in a region, the magnetic field surrounding the blood vessel is distorted slightly. This distortion shows up in an MRI image, if the imaging parameters have been properly chosen and the MRI equipment is sufficiently sensitive."

Ogawa first observed this phenomenon in animals in 1989. Experiments done over the next two years with several colleagues helped him understand the physical and neurological principles at work, and convinced him that the technique would also work on humans.

To test the idea, the Bell Labs researchers began collaborating last year with Ugurbil and his colleagues at the University of Minnesota's Center for Magnetic Resonance Research.

Minnesota had recently acquired a human-size MRI system with a much higher magnetic field strength (4 Tesla) than those commonly used today in hospitals and other radiology facilities.

This higher field strength makes possible images of sufficiently high resolution to localize brain activity to the scale of fractions of an inch, and also maximizes the contrast between hemoglobin and deoxyhemoglobin. The University of Minnesota instrument is one of only three 4 Tesla machines in the


In their experiments, the research team took an MRI image of a volunteer's brain in a relatively quiet state with no sensory stimulation, and then another with sensory stimulation--a flashing light, for example.

Image-processing software was then used to "subtract" the control image from the active-brain image, producing an image showing only the areas of increased brain activity resulting from the stimulation.

Although several other methods can be used to perform functional brain imaging, the BOLD images are unique in their combination of high-resolution, precise timing, repeatability, and safety. Some methods, such as magnetoencephalography and electroencephalography, measure the weak magnetic and electric fields produced by active brain areas.

Although these methods provide excellent time resolution, they are not able to precisely localize the source of the measured brain signals.

A more precise technique--Positron Emission Tomography (PET)--can image brain activity on the scale of centimeters, but requires much more time to acquire a functional image and, most importantly, is invasive.

It involves the injection of radioactively-labeled molecules into the subject's bloodstream, which eventually find their way to the brain. Once there, the radioactivity can be picked up by scanning devices, and provides a representation of blood flow to various regions in the brain.

Because of the necessary limitations on radiation exposure, only a small number of scans can be safely taken on an individual subject. Brain mapping in the past has therefore been done by averaging the information from several individuals, which limited the precision of the maps.

But with BOLD imaging, several consecutive images can safely be taken of a single subject, permitting studies of the onset and ending of a neural activity. In addition, BOLD images have a higher spatial resolution than PET images.