PART TWO: NEUROIMAGING
Chapter 4
Until quite recently, the power to observe a living human brain at work belonged only to mad scientists in horror movies. In the early 1990s, however, the newly developed technology of functional magnetic resonance imaging (fMRI) brought this power into the real world, and the use of fMRI in medical-scientific studies spread at an explosive rate. According to PubMed, the online citation-keeper of record, by 1985 just over two dozen research publications had discussed the prospect of developing a dynamic, real-time form of magnetic resonance imaging; by the end of 2005, the number of research publications mentioning fMRI had grown to more than 146,600.
The Power of an Image
Despite the numerous, intricate steps involved in the process of fMRI, the end product seems to offer a direct view of the cerebral scene, like the transparent plastic wall of a store-bought anthill. True, the scene from a PET scan is cluttered with information: various areas are delineated, and some specific sites bear labels; colors, ranging from black to blue, green, yellow, and red, represent different amounts of energy consumed, which in turn indicate different levels of brain activity. Nevertheless, the most powerful impression is still the simplest one, that what we are actually seeing is the brain in action. This “picture” is as persuasive as the sonogram of a thumb-sucking fetus, and just as unforgettable.
No one should be surprised that the image of an easily recognized shape, brightly lit up with attractive colors, makes a powerful visual impact. The effect may even be stronger on people who have seen very few medical images before, if any. By contrast, a trained radiologist may scarcely notice the overall picture in focusing on the specific information contained in each pixel.
“When it comes to brain imaging, most people don’t understand that they’re looking at a mathematical abstraction,” says Paul Root Wolpe, director of the Program in Psychiatry and Ethics at the University of Pennsylvania School of Medicine. “Often they also don’t understand that PET scans for research are averaged over many individuals, and even clinical scans for specific patients are often put into the frame of someone else’s brain.” Many quirks of the brain-imaging process are not apparent to the casual viewer. The images seem to be appearing more and more often these days, but with less and less explanation of their contents. Therefore, Wolpe goes on, “I think we have a particular responsibility in the public forum to make sure that people understand what an image is when they see it on television or in popular magazines.”
One important point that has been largely overlooked in the wholesale acceptance of PET imaging is that the exact results differ among individuals. The differences are apparent in response to even the simplest stimulus. For the purpose of research, though, the data from groups of study participants are often averaged together to produce a single very clear response. The first decade and a half of PET imaging made great use of the formulas for averaging, while individual differences received relatively little attention. Now, however, imagers can collect higher-resolution data, which will allow them to study individual brain responses in detail. Ideally, this new precision will lead to greater sensitivity on neuroethical issues. “With all the hype as well as the hope,” says Ruth Fischbach, director of the Center for Bioethics at the Columbia University College of Physicians and Surgeons, “we must be circumspect about emergent technologies. For many advances, there may be and indeed undoubtedly will be unintended consequences which we must confront.”
What Are We Seeing?
When brain cells do something, what are they doing, and why is it that we can see a result? One possibility is that the activity represented in a brain scan is all excitation—that is, the release and reception of glutamate at various brain sites may account for all the visual effects that we call activation. In that case, how is its opposite, inhibition— that is, the prevention of cell signaling—visually represented in the same scan? Scientists are currently designing studies that will address such questions.
Something that does not appear in a brain scan, but is important to keep in mind, is the enormous number and variety of processes all surging along at the same time. The brain accounts for only about 2 percent of the weight of the human body, but it uses about 11 percent of the output of the heart and about 20 percent of all of the energy consumed by the body—ten times the usual amount for an organ of that weight. This means that when the data from a brain scan are plotted on a graph, each data point represents a huge amount of brain activity. The irregularities in the curves on the graph, which look like background noise, may hold as much information as the curves themselves.
Individual differences in brain activity do not necessarily produce different outcomes, says Marcus Raichle, co-director of the Division of Radiological Sciences at the Washington University School of Medicine in St. Louis. For any given task assigned to study participants, brain scans may reveal a different strategy being applied by each individual; but, depending on the circumstances and the difficulty of the task, the researchers may or may not observe differences in the result.
Moreover, we cannot take it for granted that activity in the same specific area on several people’s brain scans means that these people are all experiencing the same thing. To prove that, scientists would need to carry out what are known as mechanistic studies, in which they probe a very small, defined area, make predictions of exactly what the research subjects should be experiencing, and then check the accuracy of their predictions. They must not reason back from the effect to the cause—that is, they cannot say that because they observe activity in a particular area while a study participant is making a decision, it means that that area is necessarily involved in the decision making. It can mean nothing in itself, since it is only an observation. Of course, scientists must be sure of their observations, but what the observations mean is open to interpretation. In discussing imaging with the public, they must distinguish between direct observations and their interpretations.
The individual differences are very small in relation to the overall metabolic activity of the brain. The moment-to-moment changes are tiny, not like those in a muscle of the arm or the leg when it switches from a resting state to an active state. The brain, when we are sitting quietly, is anything but at rest; neurocircuits in every region are still sending messages by the millisecond to keep us seeing, maintaining our balance, and making sense of our perceptions, all the while monitoring our environment for anything out of the ordinary that might require a new response.
A challenge that remains for researchers is designing ways to study the brain’s continuous housekeeping processes, without having it take on any additional tasks related to the study. In other words, how can we understand what the brain is doing when we’re not manipulating it?