Charles Whitman lived a fairly unremarkable life until August 1, 1966, when he murdered 16 people including his wife and mother. What transformed this 25-year-old Eagle Scout and Marine into one of modern America’s first and deadliest school shooters? His autopsy suggests one troubling explanation: Charles Whitman had a brain tumor pressing on his amygdala, a region of the brain crucial for emotion and behavioral control.

Can murder really be a symptom of brain disease? And if our brains can be hijacked so easily, do we really have free will?

Neuroscientists are shedding new light on these questions by uncovering how brain lesions can lead to criminal behavior. A recent study contains the first systematic review of 17 known cases where criminal behavior was preceded by the onset of a brain lesion. Is there one brain region consistently involved in cases of criminal behavior? No—the researchers found that the lesions were widely distributed throughout different brain regions. However, all the lesions were part of the same functional network, located on different parts of a single circuit that normally allows neurons throughout the brain to cooperate with each other on specific cognitive tasks. In an era of increasing excitement about mapping the brain’s “connectome,” this finding fits with our growing understanding of complex brain functions as residing not in discrete brain regions, but in densely connected networks of neurons spread throughout different parts of the brain.

Interestingly, the ‘criminality-associated network’ identified by the researchers is closely related to networks previously linked with moral decision making. The network is most closely associated with two specific components of moral psychology: theory of mind and value-based decision making. Theory of mind refers to the capacity to understand other people’s points of view, beliefs, and emotions. This helps you appreciate, for instance, how your actions would make another person scared or hurt. Value-based decision making refers to the ability to judge the value of specific actions or their consequences. This helps you see not only what the outcomes of your actions will be, but whether those actions and outcomes are good or bad. The letters written by Charles Whitman on the eve of his killing spree provide a chilling window into a mind losing the ability to understand good, bad, and other people: “It was after much thought that I decided to kill my wife, Kathy…I love her dearly, and she has been as fine a wife to me as any man could ever hope to have. I cannot rationally pinpoint any specific reason for doing this.”

This research raises troubling questions about Charles Whitman and the other subjects in the study—and for all of us. If their actions were caused by brain damage and a disrupted neural network, were they acting under their own free will? Should they be held morally responsible for their actions and found guilty in a court of law? Should we see them as patients or perpetrators—or both?

Some scientists have followed cases like Charles Whitman’s down the slippery slope, reaching the most extreme conclusion: that by uncovering the biological causes of behavior, neuroscience shows that “free will, as we ordinarily understand it, is an illusion”.

But these arguments depend on a faulty conception of free will. Free will should not be understood as a mysterious ability to cause actions separate from our brain activity. In fact just the opposite might be true: that free will requires certain connections between our brains and our actions. After all, our brains are the biological basis of our identity, housing our memories, our values, our imagination, our ability to reason—in other words, all the capacities necessary to make choices that are uniquely our own, and to carry out actions according to our own will.

This understanding of free will allows us to ask more sophisticated questions about the connection between the brain and criminal behavior when evaluating cases like Charles Whitman’s. Instead of just pointing to the obvious fact that an action had a neural cause (every action does!), we can ask whether a person’s specific neurologic injury impaired the psychological capacities necessary for free will—imagining possible courses of action, weighing relevant reasons, perceiving the moral features of actions and outcomes, making decisions that align with our values, and controlling behavior against competing impulses.

The specific components of moral psychology disrupted by lesions in the criminality-associated network may indeed interfere with these abilities: value-based decision making and theory of mind are important for grasping the moral impact of our actions and understanding how they will be experienced by other people. If a person has genuine impairments in these capabilities, then they possess only a diminished form of free will. Future research should evaluate more robustly the degree to which these and other psychological capacities are truly impaired in patients with lesions in this network.

When moving from the question of free will to issues of moral responsibility and legal guilt, it is important to evaluate each case in light of the wide array of factors beyond neurologic injury that influence behavior. Previous research has demonstrated that criminal behavior is impacted by genetics, childhood mistreatment, low self-esteem during adolescence, lack of parental supportsocial and economic disadvantage, and racial discrimination. Digging deeper into Charles Whitman’s case, we might wonder whether his extraordinarily strict father, or his fascination with guns as early as age 2, contributed to his later violent turn. The lesson is that human behavior is complex and a brain lesion is neither necessary nor sufficient for criminal behavior: after all, there are nearly 700,000 people living with brain tumors in the US and approximately 800,000 people have strokes every year, but the known cases leading to criminal behavior number in the dozens. Further research would be helpful in determining the likelihood that patients who suffer brain lesions in the ‘criminality-associated network’ actually go on to commit crimes, with the expectation that this kind of impairment will emerge as one of many factors increasing the risk of criminal behavior.

The fact that violence can be a symptom of brain disease shows not that free will is an illusion, but that free will can be injured just like other human abilities. These rare cases of dysfunction allow us to see more clearly that our healthy brains endow us with remarkable capacities to imagine, reason, and act freely.