Functional MRI (fMRI) is similar to a PET or SPECT scan, but it avoids radiation. This article includes a technical section foi the scientifically inclined, which you may pass over and still get the gist of the article as it pertains to pain imaging.
One of the best bets for documenting pain activity is brain scans. Functional MRI has been very limited to the present time, but there is wonderful news. Many insurers have decided fMRI is reimbursable.
This may not sound like much, but the money that will flow from this decision will force radiologists to become familiar with fMRI. Up until the present, it was largely a curiosity. No one wants to get left in the dust, so everyone will be ordering an fMRI scanner so they can advertise on TV that they are a world class hospital, there in Whereeverville, USA. Money plus prestige. You can’t beat it, and General Electric’s scanner division is jumping for joy. So are we. It will become harder and harder to ignore our story. People always give way to what they can see (as opposed to what you cannot say).
Before long, even Podunk USA will have one. The radiologists in Podunk will pretend to be able to read it; and eventually the big researchers in the university hospitals will make it a routine part of radiology residency. As this happens, they will discover what we have been telling them here–that MANY factors influence brain scans.
News of the decision to pay for fMRI has already been circulating among the neuroradiologists. Harvard has even begun offering an fMRI course for radiologists in practice, said to be the best in the country, which is doing land office business.
As to pain, this can only be good. The anatomic realities of pain will begin to trickle out of this flood of fMRIs and science will be the winner. Right now, most radiologists believe once an fMRI, always an fMRI. However, once they learn that these studies change from day to day, just as patient’s brain thoughts change from day to day, and that there will be a reason to REPEAT the very expensive studies, we will be even more welcome at the office. This will lead to a better understanding of why some CP patients have failed to show changes. We have suggested before that the mere sound of the magnet clacking is enough to change the fMRI. The realization that scanning is dynamic (changing) has been very hard to put across to radiologists who are used to a one shot deal.
Also sufficient to change things is the monotone voice that suddenly comes through some five dollar speaker into the five million dollar human pencil sharpener shaped cocoon and says, “It is very important that you hold absolutely still and do not swallow or cough for the next fifteen minutes”–a statement that never fails to make your nose itch like crazy, your throat to become incredibly irritated, and your salivary glands to shift to full throttle and fill your mouth with spit until you desperately want to swallow before you choke. There are lots of things they don’t mention that you also figure you better not do.
The initial perception of why brain scans work is that blood is shunted to the active brain area, yielding a signal as this increased volume moves to and through an area. The idea was that oxygen was being rushed to where it was needed by simple mechanisms of flow. This idea made sense, but unfortunately is now thought to be incorrect, although cerebral blood volume is still an important parameter. The scientists are much more concerned with chemical inducement (by glutamate and other neurotransmitters) of oxygen usage, and using your brain actually appears to deplete oxygen for the first moment or two. This makes more a much more complicated picture than simply looking for the blood surge. The interest now is in the shift from blood in the large vessels to blood in the tiny capillaries which represents blood going to the neurons and glia. This level of sophistication is only going to be possible with MRI machines with higher Tesla (greater levels of magnetization). Any movement of electrons induces a magnetic field. Electricians use the “right hand rule”, which means if you imagine your right hand around a live wire, the magnetic field will be directed in the direction of your thumb.
The reverse is also true. Any magnetic field will be associated with the movement of electrons. This is the principle of magnetic picture taking. What is imaged is the release of energy which the MRI machine has previously pulsed into the tissue at the calculated frequency which generates “resonance” or unique excitement of the nuclei. The frequency for medical imaging is the frequency which excites the nucleus of hydrogen atoms, but other atoms could be used.
The electromagnetic pulse puts nuclei in a higher energy state, but they quickly return to normal, releasing the energy, which then creates a field which the magnet measures and by pixellating signal pickup, a picture is created.
In an individual neuron, a firing is known as an action potential. However, brain scanning is much too broad for such measurement. Instead, scientists speak of action in the aggregate, which they call field potential. Field potential is roughly related to cerebral blood volume in a given area (see below).
fMRI usually is done using measurements of blood oxygen differences (BOLD). Blood oxygen level dependant (BOLD) images come from the fact that with neural activity, the oxygen level drops. In the past, this was thought due to simple metabolism, but now the theory is that glutamate and other neurotransmitters control oxygen utilization. The role of glutamate should make pain a suitable target for fMRI. Once the oxygen is used the BOLD signal tends to diminish or disappear because the signal from blood matches that from tissue, yielding no difference and therefore no contrast by which to make a BOLD image.
BOLD images are done in T2 mode. The BOLD signal increases as the square of the magnetic field strength in Teslas (a unit of magnetism). Existing hospital MRI scanners are 1.5 Tesla, but fMRI researchers prefer 3 Teslas or greater. The reason is that with 3 Tesla machines, the positive BOLD signal is preceded by a brief dip at about 1 second after stimulus. The positive BOLD rise occurs at about 2-5 seconds. The brief dip is felt to originate in small vessels increasing metabolism and using oxygen in response to stimulus. It is a very brief dip, but provides possibilities for improved signal tracing and localization of any abnormality.
This time factor of five seconds limits studies of ongoing pain. However, it should allow us to know where pain is in the brain and if a CP patient has signal in that area, conclusions may be drawn. At 4 Teslas and beyond, the BOLD signal stops coming from large vessels and begins coming from very small vessels (capillaries) making even greater localization possible.
Half of the hydrogen atom nuclei spin one direction and the other half spin in the opposite direction. Rememmber that MRI is done by injecting a radiofrequency into the body which “despins” or derandomizes the spin of the nuclei of some hydrogen atoms (mostly in the water in your body) into a higher energy state, ie a spin different from their original direction. The nuclei quickly “relax” or release the energy and return to their random state, but a powerful magnet can detect the energy which is released as the atoms relax into their normal spin. The magnetic flux can be turned into a picture.
______________________________________________
TECHNICAL SECTION
What is magnetic resonance, or magnetic resonance imaging (MRI)? Oversimplifying, there is an ideal frequency for the radiomagnetic pulse, calculated by what is called the Larmor formula. This can be analogized to the resonant ring which fine crystal displays when you run your finger around the rim of a crystal glass. Resonance also occurs when “exciting” the nuclei of atoms. The frequency of resonant excitation is the Larmor frequency. Energy is delivered in pulses and the magnet records the relaxing echo as the resonance creates the “ring”.
Manipulating the pulse time and strength can result in reversing the spin of a hydrogen nucleus by 180 degrees, 90 degrees, or whatever flip angle is desired. Each different kind of tissue has a different relaxation time. To minimize signal from fat or fluid, sometimes an inverting 180 degree pulse is used to stand nuclei on their head (reverse the longitudinal magnetization), then hit again with a 80 degree pulse at a time which matches the T1 relaxation time, then do the actual image revovery gradient pulse of say 180 degrees. The resulting image derived is known as inversion recovery.
Although nuclei spin according to quantum mechanics, which displays both wave and particule behavior, models usually focus on the particle or material characteristics. In that light, think of a nucleus as a golf ball tied to an elestic string above and one below. If you twist the golf ball 180 degrees, it wants to return to its original state. If you quickly twist further and then release it, its return to the relaxed state will be slightly different. Although fudging, this model allows you to conceive how nuclei behave in inversion recovery.
Remember that the relaxation emits energy which can be measured by a magnet. Remember also that the magnetization measrment of a nucleus can occur in more than one plane. Spin has only two directions, which, by convention, are called “UP” and “DOWN”. The measurement of the energy released is picked up three dimensionally in the x, y, and z planes, just like any three dimensional model, such as you learned in geometry.
A current idea for getting more information out of brain scans is to measure cerebral blood volume in a local area. Two techniques are used. The first manipulates the frequency of the pulse and the second manipulates the inversion sequence. They are known as STAR and FAIR.
STAR depends on alternating radiofrequencies by which to induce the magnetic spin in tissue nucleu, This is written LL-EPI-STAR, which stands for Look-locker echo planar imaging, signal targeting alternating radiofrequency).
By comparision, FAIR has the long name LL-EPI-FAIR, which stands for Look-locker-echo planar imaging, flow sensitive alternating inversion recovery). You don’t have to remember any of this of course, but there is likelihood pain literature of the future may refer to STAR or FAIR images. Either way, what they are after is to know how much volume of blood is in a given area at a given time.
END OF TECHNICAL SECTION
_____________________________________________________
There are already commercial machines capable of doing 7 Teslas, but this is really supermagnetic. One neuroradiologist related to this author that at 5 Teslas, the wrench which hangs from the controls, used to adjust the machine, was standing out taut horizontally on its retaining cord ten feet away from the machine while it was running a scan. We have enough problems without wrenches coming at us while we are trying to hold still.
This is reminiscent of Magneto of the X-men comic series flinging metal around. The probability that the MRI machine of the future will have high Tesla strength is one reason dentists tend to use less of the old mercury amalgam for fillings in favor of nonmagnetic oomposites, also nonmagnetic titanium for implants. Neurosurgeons not infrequently used “nonmagnetic” tantallum clips as an alternative to sutures to seal blood vessels during surgical procedure. Now, no one is entirely certain how tantalum will behave at really high Tesla strengths. Is tantalum magnetic at 7 teslas? No one is sure how implanted clips will behave.
Since spinal cord injured patients with central pain are more likely to have had neurosurgery, with possible use of tantalum vascular clips, they may be at some disadvantage in really high magnetic fields At the present 1.5 Tesla ratings, tantalum appears not to cause a problem of clip movement, although it may interfere with an optimal image. All metals have some megnetic quality, although tantalum has very little. Hopefully tantalum clips will not prove a problem, but just in case, it would be a good idea to find out from your neurosurgeon whether clips were used in any surgeries you had. This may be important when high Tesla machines become commonplace.
In an article by Gizewski et al, in Neuroimage. 2007 Jun 14, it was reported that with echo planar fMRI at 7 Tesla, the sensation of tapping a finger could be detected in all seven volunteers having fMRI, while the sensory signal was only detected in ONE volunteer at 1.5 Tesla. This message was in all cases recorded in the thalamus. Those with thalamic pain, which is a synonym for central pain, will surely benefit from future knowledge of this sort. Some centers confine the term “thalamic pain” to conditions deriving from stroke, but there is no historic requirement for this distinction. It is a matter of choice and convention at any particular institution. We are not actually certain that the thalamus is actively involved in ANY central pain. There is just as much evidence that FAILURE of the thalamus to act properly is the cause, not the generation of a pain signal.
Resolution of the fMRI, even at high Tesla, is still only about 0.5 mm. This resolution is what 1.5 Tesla machines claim but possibly cannot really deliver. Since the anterior spinothalamic tract, which carries the pain of burning, is smaller than 0.5 mm, this will make it hard to image pain as it travels through cord.
At any rate, the doors have been flung open by the insurers. This will be our first, but hopefully not our last homage to third party payers. Look for a flood of articles, and gradually developing facility in this technique. There are smart radiologists everywhere and some of them are going to begin wondering about pain. Count on it.
___________________________________________________________
Note: If you get an fMRI, do not consume any caffeine beforehand, such as cola or coffee. It blocks the BOLD signal and may explain some of the negative scans in those with central pain. Most CP patients have positive changes in at least some of the expected brain areas which process pain.
