03/18/2010 05:12 am ET | Updated Nov 17, 2011

The Emperor's New FMRI

Suppose you are asked to sit in judgment in a manslaughter case in which one man killed another in a bar fight. The facts are not in dispute. The issue is the responsibility of the perpetrator. The defense calls an expert witness, a brain scientist, who reports that the defendant has a brain tumor in a particular part of his brain, and that 50% of people with such tumors commit serious acts of violence. Do you hold the defendant responsible? Or, in an alternative scenario, the defense calls an expert witness, a psychologist, who reports that the defendant was abused by his father as a child, and that 50% of men with histories of abuse commit serious acts of violence? Do you hold the defendant responsible?

When John Monterosso, Ed Royzman, and I did a study very much like this a few years ago, we found that if the defendant had a brain tumor, people were likely to absolve him of responsibility, even if the expert witness testified that only 20% of such people do violent deeds. In contrast, if the defendant suffered abuse, people were likely to hold him accountable, even if the expert witness testified that 90% of such people do violent deeds.

Monterosso, Royzman and I interpreted this result to mean that for lay people, brain states cause behavior, so that a person with a brain defect could no more help doing what he did than a dropped glass can help falling to the floor. "The tumor made me do it." In contrast, psychological states do not cause; they may suggest action, but they do not compel it. The defendant, abused childhood or no, could always have resisted the temptation to be violent.

This lay attitude, which we termed "naïve reductionism," assumes that the causes of behavior lie in states of the brain, not in states of psychology. This is a profound mistake. It rests on a picture of the world, known as dualism, that we have inherited from Descartes. For Descartes, there is body, and there is mind, and we can have scientific laws about bodies, but not about minds. Nowadays, there are not many people, and certainly no social or neuroscientists, who take this view seriously, at least not when they are being reflective. All states of mind simply are states of the brain. There is no other "stuff" but material stuff. So when a psychologist testifies about the relation between childhood abuse and violence, the testimony is that there is a reliable, causal relation between one and the other. No doubt, the brain of an adult who was abused as a child is different from the brain of an adult spared abuse as a child. But we don't need to know anything about those differences to say something meaningful about how childhood abuse causes violence. As I said, we all know this when we're being thoughtful. But when we're not paying attention, we tend to lapse into the Cartesian dualism of old.

I mention this study of ours because of the great enthusiasm currently being manifested by the lay public for "affective neuroscience." Put people in FMRI scanners, show them stuff, and see what parts of the brain "light up." What have these studies taught us? We now know that when Yankee fans see a Yankee win, their "reward centers" are active. When they see a neutral team win, there is no such activity. When people find out their stock portfolio has gone up, the same parts of their brain light up as when they are eating a delicious meal. Amazing stuff! Without this research, we might not take seriously the enthusiasm fans show for their sports teams or the avidity with which people track stock market performance. Does this affective neuroscience tell us something we don't already know? The reason we may think so is that the Cartesian dualist in us distinguishes "mere" psychology from neuroscience, which is, after all "science."

Let me hasten to say that not all of this new affective neuroscience affirms the obvious, and it may turn out that as the science develops, it may teach us many important things that we wouldn't otherwise know. Some of it already has. Neuroscientist Joseph LeDoux, for example, has traced out two quite different pathways involved in fear responses, one of which goes through the cortex, and thus involves interpretation and evaluation, and the other of which does not. This suggests that there are reflective and intuitive fear responses, a distinction that might not have been apparent by just watching our behavior. So affective neuroscience has a lot of potential. But the only reason I can see for being so excited by much of what affective neuroscience has shown us thus far is the naïve reductionist view that if the brain is doing it, it's real, and it's science. It is real and it is science whether or not we know how the brain is doing it because we know, whenever we are happy, sad, afraid, guilty, angry, or anything else, the brain is doing something. We should save our enthusiasm for findings that actually move our understanding forward and escape once and for all the dualist trap that Descartes has ensnared us in. And journalists should avoid the mind candy of brains lighting up and help us think more carefully and critically about what counts as scientific progress and what does not.