THINKING CRITICALLY What Can fMRI Tell Us About Behavior and Mental Processes

58
Chapter 2 Biology and Behavior
FIGURE 2 .7
Combining a PET Scan and
Magnetic Resonance Imaging
Researchers have superimposed images
from PET scans and MRI to construct a
three-dimensional view of the living brain.
This figure shows the brain of a young
epileptic girl. The picture of the outer
surface of the brain is from the MRI; the
pink area is from the PET scan and shows
the source of epileptic activity. To the right
of the figure are separate MRI and PET
images taken at one plane, or “slice,”
through the brain (indicated by the line on
the brain at the left).
A
T H I N K I N G C R I T I C A L LY
picture may be worth a thousand
words, but the pictures of brain activWhat Can fMRI Tell Us About
ity offered by fMRI are generating
Behavior and Mental
millions of them. As of 2005, more than 4,000
scientific articles have reported the results of
Processes?
fMRI scans taken while people engaged in
various kinds of thinking or experienced various emotions. Neuroscientists who use
brain imaging techniques are now to be found in psychology departments around the
world, and, as described in other chapters, their work is changing the research landscape in cognitive, social, and abnormal psychology. Excitement over fMRI is not confined to scientists, however. Popular and scientific magazines routinely carry fMRI pictures that appear to “show” people’s thoughts and feelings as they happen.
■ What am I being asked to believe or accept?
In the early 1800s, similar excitement surrounded phrenology, a technique that involved
feeling bumps and depressions on the skull. It was claimed that these contours
reflected the size of 27 structures on the brain’s surface that determine personality
traits, mental abilities, talents, and other characteristics. Although wildly popular with
the public (Benjamin & Baker, 2004), phrenology did not survive the critical thinking
of nineteenth-century scientists, and the technique has long been discredited. Today,
some scientists wonder whether fMRI is a twenty-first-century version of phrenology,
at least in the sense that their colleagues might be accepting its value too readily. These
scientists point out that, although fMRI images can indicate where brain activity occurs
as people think and experience emotion, there is no guarantee that this activity is actually causing the associated thoughts and feelings (Aldridge, 2005). Questions are also
being raised about the assumption that particular thought processes or emotions occur
in a particular brain structure, or set of structures. It is easy to talk about “thinking” or
“attention,” but these psychological terms might not correspond to specific biological
processes that can be isolated and located by any technology. In short, critics claim that
the results of fMRI scans can be misleading and that they don’t necessarily tell us much
about how the mind works (Uttal, 2004). Perhaps it would be better to focus on how
the brain produces thoughts and feelings instead of searching for their locations.
■ Is there evidence available to support the claim?
When the participant in an fRMI experiment thinks or feels something, you can actually see the colors in the brain scan change, much like the color changes you see on
The Central Nervous System: Making Sense of the World
59
weather radar as a rainstorm intensifies or weakens. Looking at an fMRI scan, you get
a clear impression that the brain areas that “light up” when a person experiences an
emotion or performs a mental task are the ones involved in that emotion or task (see
Figure 1.5).
These scans are not as precise as they seem, though, because fMRI doesn’t directly
measure brain cell activity. The colors seen in an fMRI scan reflect instead the flow of
blood in the brain and the amount of oxygen the blood is carrying. Changes in blood
flow and blood oxygen are related to changes in the firing rates of neurons, but the
relationship is complex and not yet fully understood (Buxton et al., 2004). Further,
when brain cells process information, their firing rates may either increase or decrease
(Gonsalves et al., 2005). If the increases and decreases in a particular brain region happen to cancel each other out, an fMRI scan will miss the neuronal activity taking place
in that region. In fact, compared with the direct measurement of brain cell activity that
can be done in research with animals, fMRI technology is still rather crude. It takes
coordinated changes in millions of neurons to produce a detectible change in the fMRI
signal.
Critics also argue that the results of fMRI research can depend too much on how
experimenters choose to interpret them. In a typical fMRI experiment, participants
are shown some kind of display, such as pairs of photos, and asked to perform various tasks. One task might be to press a button if the photos are exactly the same. A
second task might be to press the button if objects in the photos are arranged in the
same way. In this second task, a participant should press the button if one photo
shows, say, a short man standing to the left of a tall woman and the other photo shows
a small dog standing to the left of a giraffe. Both versions of the task require the participant to compare two images, but only the second of them requires considering
whether things that look different are actually similar in some way. The fMRI scans
taken during these tasks might show certain brain areas “lighting up” only during the
second task. If so, the researcher would suggest that those areas are involved in recognizing analogies, or the similarities between apparently different things (Wharton
et al., 2000). The researcher would base this conclusion on a computer program that
compares fMRI scans taken during two tasks, subtracts all the “lighted” areas that are
the same in both scans, and keeps only those that are different. But what the computer classifies as “different” depends on a rule that is set by the experimenter. If the
experimenter programs the computer to display only big differences between the scans,
not many “lit up” areas will remain after the comparison process. But if even tiny differences are allowed to count as “different,” many more “lighted” areas will remain
after the subtraction process. In our example, then, there could be large or small areas
apparently associated with recognizing analogies, all depending on a rule set by the
researcher.
These problems aside, critics wonder what it really means when fMRI research shows
that certain brain areas appear activated during certain kinds of tasks or experiences.
Their concern focuses on studies such as one from the new field of neuroeconomics that
suggests that excessive activity in a particular brain area leads to bad investment decisions (Kuhnen & Knutson, 2005). Another fMRI study claimed to show the “neural
basis of romantic love.” In this study, investigators scanned people’s brains as they
looked at pictures of their romantic partners and compared these scans to those taken
while the same people viewed nonromantic friends (Bartels & Zeki, 2000). According
to the “difference” rule established by the experimenters, four brain areas were more
active when viewing a romantic loved one than when viewing a friend. But does this
result tell us anything about how or why these areas became active or what results this
activity might have? In other words, do we know more about love? Critics of fMRI
would say no.
■ Can that evidence be interpreted another way?
Supporters of fMRI disagree. They believe that the colorful areas seen on fMRI scans
can provide vital new information that will eventually allow scientists to answer
60
Chapter 2 Biology and Behavior
EXPLORING BRAIN FUNCTIONS
WITH fMRI As this participant performs
a mental task, a functional magnetic resonance imaging scanner records blood flow
and blood oxygen levels in her brain. The resulting computer analysis shows as “lit up”
areas the parts of the brain that appear to
be activated during the task, but critics
doubt that fMRI scanning is as clear or
accurate as its proponents suggest.
important questions about behavior and mental processes. They point, for example,
to fMRI research on brain mechanisms that help us to appreciate what other people
are feeling—that is, to experience empathy—and to learn by watching others.
These mirror neuron mechanisms were discovered accidentally by scientists who
had been using surgical techniques to directly record the activity of brain cells in
monkeys’ brains (Rizzolatti et al., 1996). They found that neurons in an area called
F5 are activated not only when a monkey plans to reach for an object, such as a
peanut, but also if the monkey sees an experimenter reach for a peanut! After fMRI
scanning became available, researchers could begin looking for mirror mechanisms
in the human brain. And, in fact, some of the mirror systems they found in humans
correspond to the F5 region in monkeys (Rizolatti & Arbib, 1998). One of them is
called Broca’s area, and, as described later, it is an important component of our ability to speak. It makes sense that Broca’s area contains a mirror mechanism, because
language is a skill that we learn partly by imitation. The new fMRI findings suggest
that Broca’s area may also be important for many other skills that involve imitation.
One recent study found that this area “lights up” when a guitar student learns chords
by watching a professional guitarist (Buccino et al., 2004). Other fMRI research
has found that mirror systems in other parts of the brain become active when a
person sees someone experiencing emotion. For example, the brain area that is activated when you experience disgust (from the smell of rotten eggs, for example) is
also activated if you see a video in which someone else reacts to a smell with disgust
(Wicker et al., 2003).
So fMRI can be uniquely useful, say its defenders. Without it, research on mirror
neurons in humans could not have taken place. And because of it, we have evidence
that the experience of empathy comes about because seeing the actions and emotions
of others activates the same brain regions that would be active if we were doing
or feeling the same things ourselves. Some fMRI studies have also found that malfunctioning mirror mechanisms are associated with the impairments in language
development, in imitative skills, and in empathy seen in children diagnosed with
autistic disorder (Dapretto et al., 2006; Miller, 2005b; see the chapter on psychological disorders).