This article arises from Future Tense, a partnership of Slate, the New America Foundation, and Arizona State University that examines emerging technologies and their implications for society. On Monday, Oct. 22, Future Tense will host “My Brain Made Me Do It,” an event in Washington, D.C., on how neuroscience advances are affecting the judicial system. To learn more, to RSVP, or to watch the live webcast, visit the New America Foundation website.

It was 30 minutes into the first day of the year 2000, no dire Y2K computer failures had materialized, and Andrew was partying on the streets of Las Vegas. He had been drinking, of course, as had all the girls swarming around. The girls were dressed to be noticed and didn’t seem to mind much when the men groped them in passing.

So Andrew probably wasn’t expecting a 5-foot-tall blond girl to take offense at being grabbed. She spun around and glared at him. Andrew grinned—and the girl slapped him. In response, his fist curled and his arm hauled back. Soon the girl lay sprawled on the sidewalk, nose broken and seeping blood, bruises blooming under her eyes. There were cops around, but Andrew escaped into the crowd before they could respond.

Rage attacks are violent outbursts grossly disproportionate to the circumstances. Information about sources of frustration or irritation—like the sting of a slap—converge upon a primitive structure nestled deep in the brain stem called the periaqueductal gray. This region is responsible for coordinating essential evolutionary behaviors, from nursing young to self-defense to aggressive attacks. Information from this structure is fed upward to the amygdala and hypothalamus, which refine and coordinate body responses that accompany moments of fury. Among people who are particularly emotionally volatile, surges of violent rage orchestrated by this circuit can be especially dramatic.

Although the behaviors associated with rage attacks—carefully aimed blows, perhaps the use of a weapon—seem as though they must be consciously coordinated and deliberate, this is an illusion. Even complex behaviors that emerge during rage attacks can result from simple electrical stimulation to any of the primitive structures that make up the rage circuit. A jolt of electricity to the right part of the hypothalamus can turn a placid cat into a hissing, clawing demon, attacking nearly any creature within reach. Turn off the current, and the cat returns to its placid state.

Imagine your own cat attacked you during an electrically induced rage attack. Would you blame Fluffy for her behavior? Would you punish her?

These seem like absurd questions until we consider cases of humans whose rage attacks result not from electrodes implanted in the hypothalamus or amygdala, but from tumors or cysts in these regions. In one famous case, a man named Mark Larribus attacked and nearly killed his girlfriend’s young daughter after her crying sent him into a fit of fury. A psychiatrist determined that the frequency and severity of his rage attacks had been increasing over several months and ordered a CAT scan, which showed a tumor compressing his amygdala and hypothalamus. The tumor was removed, and Larribus’ rage attacks, which he described as feeling as uncontrollable as a car skidding across a patch of ice, ceased, and he was cleared of all charges. He was not blamed or punished. If you think this is a fair outcome, you probably believe that it is unfair to punish someone for even heinous violence if that violence results from a clear brain abnormality.

Larribus’ case was a fairly simple one—the biological trigger for his violent behavior was easily identifiable and reversible. The same cannot be said for the dozens of other biological anomalies linked to violence that have been discovered during the upsurge in human neuroscience research over the last several decades. Among the most well-known is a variant of a gene called MAOA, which is linked to extreme emotional volatility and violence. Particularly among males maltreated during childhood, possession of the risky variant of this gene increases the odds that a person will commit a violent crime during his lifetime. Brain imaging studies find structural and functional changes in the amygdala and other brain regions among men who carry this risky gene.

The case of Mark Larribus suggests that we hesitate to imprison individuals whose violence results from causes beyond their control, and for whom punishment will not improve their behavior—especially if genuinely effective treatments exist.

But what about criminals whose disorders can’t be treated, like psychopaths? As many as half of all violent offenders may be psychopaths, meaning they show little empathy or remorse and are likely to reoffend. Normally a diagnosis of psychopathy is an aggravating factor that results in a longer sentence. But emphasizing the biological basis of the disorder can reduce even psychopaths’ sentences. In a recent study published in Science magazine, trial judges presented with a hypothetical situation reduced the recommended sentence for a convicted psychopath by more than a year after they read evidence linking psychopathy to atypical MAOA activity and amygdala function. Judges justified their decisions by explaining that even though the convict was likely to remain dangerous, he was mentally ill and could not control his actions—that he was, in a sense “disabled.” But unlike the case of Mark Larribus, reducing this offender’s sentence would result in the early release of an untreated, violent man.

Another troubling feature of this study was that it presented the risky MAOA allele as contributing to the development of psychopathy. In fact, research suggests that these two phenomena increase violence via opposite mechanisms. The risky MAOA gene increases aggression by making carriers over-reactive to emotional events like frustration or threat, whereas psychopaths are predominantly emotionally under-reactive, showing little anxiety or fear even when threatened. That biomechanical evidence inconsistent with the current understanding of psychopathy still swayed judges’ opinions highlights the need for extreme caution when introducing biological evidence into courts of law, which may not be well-equipped to evaluate the validity of scientific explanations.

And what about the victims?

I can speak as one of them—I was the girl whose nose was broken on New Year’s Day in Las Vegas, and Andrew is a pseudonym I’ve given my attacker. Only a tiny fraction of men would have reacted as he did that night, and I’m confident that “Andrew’s” brain was somehow compromised. Does he deserve blame for what he did? Punishment? Part of me thinks so, but I’m aware that retributive urges spring from the same primitive, not-entirely-rational brain structures that drive rage attacks. The better part of me hopes that the research that I and other neuroscientists are undertaking will yield not just a better understanding of these disorders that may lead to clemency in court, but, much more importantly, improved methods for prevention and treatment that will make punishing certain violent offenders as incongruous and unnecessary as punishing a cat with an electrode, or a man with a tumor.