Murderous Minds (21 page)

Many neuroscientists believe a part of the brain called the anterior cingulate cortex (ACC) plays a role in such impulsive behavior. More specifically, this part of the cortex seems to be active when the brain is making a choice, detecting errors, monitoring conflicts, and learning what to avoid.

Figure 10 shows the location of the ACC in the frontal lobe, tucked beneath other folded layers of the cortex located closer to the inner surface of the skull. The ACC is also highlighted in the MRI image in Figure 12, which makes the cortex look a bit like a maze. (This might be an apt analogy for anyone entering the brain with the intention of coming out with answers about philosophical questions related to some of its most intriguing and mysterious functions such as consciousness, free will, self-awareness, and the nature of good and evil.)

The folds that make up the cortical maze allow much more brain tissue to fit into the limited space inside the skull. Since the skull could not expand as brains evolved and still safely pass through mammalian birth canals, the evolutionary solution was to crumple up the expanding cortex like a sheet of paper scrunched into a ball. If you could spread out the cortex, it would cover an area about 2½ square feet, about the size of a sheet of newspaper, and between only 1/16 and 1/6 of an inch thick.

In 2013, a team of researchers including Eyal Aharoni, Ph.D., Michael S. Gazzaniga, Ph.D., and Kent Kiehl, Ph.D., announced that they had detected something promising in the folded cortex of the ACC. They published the results of a preliminary study called “Neuroprediction of Future Rearrest” in the Proceedings of the National Academy of Sciences. The
inspiration for the report’s title was not a story by Philip K. Dick: the inspiration was the result of a brain-imaging study involving 96 prisoners between the ages of 20 and 52. The images suggested to the authors “a potential neurocognitive biomarker for persistent antisocial behavior.”
¹⁴

A press release prepared by the Mind Research Network in Albuquerque, N.M. and released by Duke University began with the claim: “A new study conducted by The Mind Research Network in Albuquerque, N.M., shows that neuroimaging data can predict the likelihood of whether a criminal will reoffend following release from prison.” Its headline was more qualified: “Brain scans might predict future criminal behavior [emphases added].”
¹⁵

The findings had “incredibly significant ramifications for the future of how our society deals with criminal justice and offenders,” Kiehl said in the statement. “Not only does this study give us a tool to predict which criminals may reoffend and which ones will not reoffend, it also provides a path forward for steering offenders into more effective targeted therapies to reduce the risk of future criminal activity.”

PsychCentral’s headline was as optimistically circumspect: “Brain Scans Could Predict Future Criminal Behavior” but the opening was a bit more conservative: “Neuroimaging data could help researchers predict whether a criminal will break the law again once released from prison [emphases added].”
¹⁶

Kiehl, a hardworking and ambitious researcher, has gone far since getting his Ph.D. in psychology and neuroscience at the University of British Columbia in 2000. One of his goals is to discover an effective treatment for psychopathy, which no one has yet managed to do. “If you could target the brain region involved, then maybe you could find a drug that treats that region,” he told John Seabrook in a 2008 New Yorker profile. “If you could treat just five per cent of them [psychopaths], that would be a Nobel Prize right there.”
¹⁷

In graduate school, Kiehl’s mentor was Robert Hare, the pioneer psychopathy researcher and developer of the Psychopathy Checklist (PCL–R). By 2013 at age 43, Kiehl was a tenured professor of psychology with secondary appointments as professor of law and professor of neuroscience at the University of New Mexico. He is also the Executive Science
Officer and Director of the Mobile Imaging Core and Clinical Cognitive Neuroscience at the Mind Research Network in Albuquerque. The latter appointment provides him with a mobile fMRI brain-scanning machine, which he helped design.

fMRI Delivered to Your Prison

The mobile brain scanner is a crucial tool in Kiehl’s work. It gives him an advantage over researchers without an fMRI machine-on-wheels. Most of the multi-ton, multi-million-dollar machines that take snapshots of living brains are confined to hospitals or research facilities. For most studies, criminal volunteers like Willem Boerema, shackled and under guard, are transported to the machines to have their brains scanned. This is done with significant inconvenience to administrators, researchers, and guards.

Kiehl’s brain scanner goes to the criminals. Kiehl and his collaborators use a semitrailer to transport the mobile fMRI machine to the prisons and adolescent correctional facilities of the New Mexico Corrections Department. So far, they have taken more than two thousand brain scans of volunteer prisoners, many of whom are psychopaths or minors whose behavior suggests they could someday develop into psychopaths. Nobody has more scans of criminal brains than Kiehl and his team of forty investigators and staff at the Mind Research Network.

The scans that Kiehl and his co-authors claim hold the promise of predicting re-arrests indicate to the scientists that something is not quite right in the ACC, the anterior cingulate cortex. This is just what you might expect, since the ACC is one of the regions reported to show increased activity whenever someone makes a decision involving an impulsive behavior.

For example, let’s say you agree to look at a screen and press a button every time the letter “X” appears but not when the letter “K” appears. It sounds simple, but it takes concentration because a “K” is just a modified “X” with its left limbs pointing straight up and down as if it were performing a half jumping jack. Since the letter “X” appears more than 80 percent of the time in this test, it takes some effort to not press the button when “K” pops up on the screen. An impulsive person would be expected to make more errors and hit the button more often when “Ks” appear than a less impulsive person. If someone has trouble controlling the impulse to
respond when a letter pops onto the screen, scientists figure this trait might be reflected in a brain structure like the ACC which has previously been implicated in this type of task.

This is the test Kiehl and his colleagues asked their convicted volunteers in two New Mexico prisons to take while the prisoners underwent brain scans in the Mind Research Network’s mobile scanner. After following these volunteers for four years, the scientists report that those with relatively low activity in their ACCs during the test were roughly two times more likely to get re-arrested after being released from prison than were convicts with relatively active ACCs. Kiehl said “This means we can see on an MRI a part of the brain that might not be working correctly—giving us a look into who is more likely to demonstrate impulsive and antisocial behavior that leads to re-arrest.”
¹⁸

A co-author of the report, Walter Sinnott-Armstrong, Ph.D., a Professor of Philosophy at Duke University, believes that “These results point the way toward a promising method of neuroprediction with great practical potential in the legal system. Much more work needs to be done, but this line of research could help to make our criminal justice system more effective.”
¹⁹

Not surprisingly, this paper received more press coverage than most journal articles, which receive none. Many accounts in the mainstream media repeated the information in the press release. Bloggers tended to be more skeptical. Many wondered if spotting decreased blood flow to a particular region of the brain could really predict who will commit future crimes, or, more accurately, who will get caught and re-arrested for committing crimes in the future. Some neuroscientists doubted if low activity in the ACC could be a useful marker for antisocial behavior.

Kiehl has found that the ACC shows the greatest change in activity when someone is trying to control their impulses. That is why he zeroed in on this region when he and collaborators designed the experiment. Other parts of the brain suspected of being involved in controlling impulses include the dorsolateral prefrontal cortex (Figure 8) and the basal ganglia (Figure 11). The ACC may not be the main player in this type of mental activity. It is more likely part of a network and it could be responding with increased activity to subtle inputs from other brain regions.

One model presented in 2002, for instance, includes the ACC as a control filter in a network or circuit that processes information about errors.
²⁰
If this model is accurate, the ACC may receive information from the amygdala, basal ganglia, and parts of the frontal cortex and pass it along to the motor cortex (Figure 8) and other regions involved in movement. These motor areas could feed updated information back to the frontal cortex and basal ganglia. This scheme outlines a wiring diagram that might underlie one way we learn by making mistakes.

Such interconnectedness does not contradict Kiehl’s assertion that “we can see on an MRI a part of the brain that might not be working correctly—giving us a look into who is more likely to demonstrate impulsive and antisocial behavior that leads to re-arrest.”
²¹
But, just because it might not be working correctly doesn’t mean it is structurally impaired either. One comparison of a couple of dozen psychopathic subjects and an equal number of non-psychopathic subjects found no group differences in the size or volume of the ACC.
²²
It is possible that the ACC itself is not impaired in psychopaths. The ACC reduced activity in Kiehl’s and others research might reflect reduced input from other parts of the brain, a possibility that still leaves the ACC as a potentially useful marker for defective impulse control in affected individuals.

Kiehl someday hopes to capitalize on the apparent involvement of the ACC in impulsive decision-making by developing a way to train prisoners to improve its function and so improve their ability to control their impulses. This, most of us would agree, would be preferable to seeing a criminal psychopath who left his wallet at home look around for a heavy piece of wood, as the psychopath who robbed the gas station did on the spur of the moment when he suddenly found himself out of cash.

Of course, in such a controversial field of research, not everyone agrees with the statements in the press release.

“The problem with this work is that it fails to consider core principles of psychological measurement,” Christopher Patrick said.
²³
One of the most basic of these is that the brain response during the test (psychologists would call it the predictor or independent variable) must be reliable if the prediction is to be valid or useful. This requires proof that the test yields
consistent scores for individuals across time. The result of one scan is not meaningful in scientific terms until it has been demonstrated to be reliable. Considering these and other basic principles of psychological testing, “the likelihood that recidivism can be predicted from ACC activation within a single-session lab task in any way that would be practically useful is extremely low,” Patrick concluded.

Any brain scan or test used to predict the future behavior of an individual has to meet rigorous criteria to satisfy skeptical psychologists like Patrick. To ensure that the test results are meaningful and accurate, it is also necessary to take into account the limitations of the technology. Craig Bennett has learned a lot about these issues based on his investigation of how emotional information is processed and integrated with higher cognitive processes in the brain. His research combines his interests in developmental cognitive neuroscience and magnetic resonance imaging methods.

“While statistical classifiers can sometimes pick up on activity levels or patterns of activity that might be predictive, there is still a lot about how individuals vary that we can’t explain,” he said. “For example, if we gather data from a group of 200 prisoners and generate a classifier that can predict psychopathy, how do we know that prisoner #201 is really a psychopath if our classifier says he is? Could there be other, mitigating activity in other parts of the brain that our classifier is not trained on? Is 200 people really a big enough sample size to extrapolate to all people who are and are not psychopaths? Have we captured all possible variations in regional brain activity with 200 people? These are sticky questions without good answers.”
²⁴

Bennett’s comments were not specifically made in response to the claims made in the neuroprediction paper but rather directed at anyone who views or evaluates fMRI results and claims. As mentioned previously, most colorful fMRI images are combined group results placed on a single representative image of a brain. Making the leap from that group-accurate image to a single person’s fMRI scanning result may not be as easy as people outside the field assume.

“If you compare the group fMRI results to those of the individuals that comprise the group, you will typically find that they look almost nothing alike,” Bennett explained. “While the group map detects similarities that
are generally consistent across the group, there is a wide range of variation in terms of the individual differences in regional brain activity.”

Jim Fallon agrees but adds that part of the problem is not carefully defining the person you are studying. He used schizophrenia as an example. Once you diagnose someone correctly and know something about their genome, he has found, “you can get a pretty consistent look to those scans.”
²⁵
The same would apply to a population of carefully defined and diagnosed psychopaths.