Inside the Dangerous Mission to Understand What Makes Extremists Tick—and How to Change Their Minds

On a cool winter’s day in early 2014, the American academic Nafees Hamid was invited for tea at the second-story at the Barcelona apartment of a young Moroccan man. It started well enough; they sat down at the kitchen table, chatting amiably in French while two acquaintances of the host sat nearby in the living room. Halfway through the conversation, though, things took a turn. “He started saying things like, ‘Why should we trust any Westerner?’” Hamid recalls. “‘Why would we not kill every one of them? Why should I even trust you—you are an American—sitting here? Why should I even let you out of my apartment?’” The man briefly left the kitchen and went into the living room to speak to the others in Arabic, a language in which Hamid is not fluent. But he repeatedly heard one word he did know: munafiq—a term that, at best, means hypocrite; at worst, “enemy of Islam.”

“I realized that they were talking about me, and that this was going in the wrong direction,” says Hamid, who had arrived hoping to coax the Moroccan to participate in a study.

As quietly as possible, he opened the second-story window and jumped out, his fall cushioned by the awning of a fruit stand below. Adrenaline spiking, he bolted to the safety of a crowded train station a few blocks away.

Field research on jihad has its hazards. Hamid, now 36, had come to the apartment knowing—from a questionnaire he had already filled out—that the Moroccan man harbored extremist inclinations. The effort was part of a larger project to discover the roots of radicalization and what might cause someone to fight or die—or kill—for their beliefs.

But the work goes on, a part of a larger undertaking by an unusual network of policy experts and international scientists, many of whom have their own harrowing tales of escaping danger or navigating dicey situations in pursuit of groundbreaking research. Recently, the group published the first brain-imaging studies on radicalized men and young adults susceptible to radicalization. The private research firm behind the group’s work, Artis International, is officially headquartered in Scottsdale, Ariz., but doesn’t truly have a base. Its academics and analysts operate from far-flung places, tapping an array of funding from various governments, the U.S. military and academic institutions. The central goal of the firm is to advance peace by figuring out what motivates people to become violent—and how to reorient them toward conflict resolution, or prevent them from becoming violent in the first place.

Read More: This Researcher Juggled Five Different Identities to Go Undercover With Far-Right and Islamist Extremists. Here’s What She Found

That means getting as close to the perpetrators and their supporters as possible. Much of Artis’ work has been rooted in behavioral sciences and informed by straightforward research methods, like surveys. But Artis researchers have also pushed the boundaries of social science, through everything from experimental surveys on armed forces to psychological tests on imprisoned extremists. Its investigations have led researchers to the front lines of the war against ISIS, restive areas in North Africa, and lately into Eastern Europe and cyberspace.

Even by Artis standards, the recent brain-imaging studies conducted in Barcelona—the work that had Hamid leaping from a window—were remarkable for the level of risk the researchers undertook. The scientists wanted to find hard neurological evidence to support previous social-science findings and widely held assumptions: that extremists could be influenced by their peers, and later, that social exclusion may harden the beliefs of a budding extremist. To gather this sort of information, researchers like Hamid would have to scour the streets of Barcelona for extremists; somehow convince hundreds of them to take surveys; and then, after identifying the most radicalized, coax them to undergo multiple brain scans at a seaside hospital campus. What could possibly go wrong?

Origins: A Research Void

The roots of the Barcelona brain studies go back to 2005, when the U.S. government was still absorbing the 9/11 attacks. Richard Davis, who would go on to co-found Artis International two years later, had recently started working as a policy adviser for the U.S. Homeland Security Council (which reports to the President) and was alarmed by how the government came to its counter-terrorism strategies. “It became clear that many of the decisions that were being made—grand decisions about terrorism—were being made with little to no field-based scientific evidence backing them,” he says.

One key problem is that empirical extremism studies require access to materials that governments might not want to share, like transcripts of intercepted communications or interrogations, explains Liesbeth van der Heide, a research fellow at the International Centre for Counter-Terrorism in the Hague. Ideally, the studies also involve access to extremists themselves, who are even harder to come by. “There aren’t many of them,” she says. And the ones that succeed in carrying out violent plans “tend to die in an attack or flee.”

So most terrorism research has tended to draw on secondary sources—reports in the media, for example, or other books or articles already published on the subject, resulting, she says, in “an echo chamber repeating what others have said.” An exhaustive 2006 review of 6,041 peer-reviewed studies on terrorism published from 1971 to 2003 found that only 3% were based on empirical data. “Thought pieces”—articles where authors discussed an issue theoretically or offered an opinion—accounted for 96%.

This alarmed Davis. He believed that any government interested in curbing violence needed not more thought pieces, but a more scientific understanding of the people who commit it based on primary sources. Academics already doing this sort of work were rare exceptions, but both Marc Sageman, a former CIA case officer turned forensic and clinical psychiatrist, and Scott Atran, an anthropologist, had spent extensive time with members of militant jihadist groups, from the Afghan mujahedin to al-Qaeda. Davis sought them out in the fall of 2005, and by 2007 had convinced them to help him launch a firm dedicated to on-the-ground research into violence reduction. They named it Artis, Latin for “of art,” “of skill” or, in some usages, “of science.”

That same year, Artis cobbled together funding from a range of institutions—including the U.S. Air Force Office of Scientific Research, the U.S. National Science Foundation and the French National Centre for Scientific Research—to study the underlying causes of political violence. They decided to focus on a social-psychology concept called “sacred values”—a person’s deepest, most nonnegotiable values—which would lay the groundwork for their Barcelona brain scans.

Sacred Values

In the 1990s, social psychologists Jonathan Baron at the University of Pennsylvania and Philip Tetlock at the University of California, Berkeley, developed the concept of “sacred values” to counter economic theories that suggested everything had a price. Certain values (like human life, justice, civil liberties, environmental or religious devotion) could be so sacred to people that they would be unwilling to act against them, no matter the cost or consequence.

Atran, who had been studying values for decades through the lens of anthropology, began applying this concept to the study of violent extremists after 9/11. It occurred to him then that, perhaps, the perpetrators had committed the suicide attacks in defense of deep values the rest of the world had been overlooking. By 2007, Atran had advanced this line of thinking in several articles about jihadist terrorists. His Artis colleagues found evidence that material incentives may backfire when adversaries see the issues at the heart of a dispute (like land and nationhood) as “sacred.”

The Artis team continued to hone the connection between sacred values and violence into 2014, when a comment from President Barack Obama’s Director of National Intelligence, James Clapper Jr., gave them a renewed sense of purpose. In an interview, Clapper said that the U.S. had underestimated ISIS militants because predicting a group’s will to fight was “an imponderable.” In response to that comment, Atran and his colleagues decided to use their knowledge of sacred values to measure militants’ will to fight, which they believed was indeed “ponderable.”

That same year, they did survey-based research on networks in Spain and Morocco responsible for the 2004 Madrid bombings. It found that people were more willing to sacrifice their lives if they were part of a close-knit group that shared their sacred values. They also began laying the groundwork for a separate study, eventually published in 2017, that found that among members of various forces who fought against ISIS, those who expressed the most willingness to fight and die for abstract values like nationhood, heritage and religion tended to prioritize those values over their social groups, like family.

Still, by 2014 most such work had come from what fighters said in interviews or surveys. Atran was convinced that sacred values were so deep and powerful that the brain must process them differently than it processes decisions about more mundane issues. But to truly understand the relationships between neural pathways associated with such values and willingness to sacrifice for them, Atran and his colleagues believed they needed to get a look inside extremists’ heads.

Recruitment

Barcelona’s Raval district is a maze of graffiti-sprayed buildings and narrow streets. In recent years, chic galleries and boutique clothing stores have begun to spring up between halal butchers and Arabic-language bookshops, filling the boarded-up storefronts emptied by the waves of evictions that ravaged the primarily immigrant neighborhood following the 2008 financial crisis.

The locale has also been the epicenter for a number of foiled terrorist plots, and is carefully monitored by both Spanish and international intelligence bodies for jihadist activity. That made it an appealing place for Hamid and his colleagues to recruit radicalized men for their inaugural brain study on extremists. The Artis researchers planned to use a combination of behavioral tests and brain scans in a functional magnetic resonance imaging (fMRI) machine to see whether a hardened extremist’s “will to fight” for his sacred values was susceptible to peer influence.

In early 2014, the group decided to target a small pocket of extremists in Barcelona’s Pakistani community that authorities had been tracking for years. They set their sights on 20- to 30-something first-generation Pakistani men who openly supported Lashkar-e-Taiba, an al-Qaeda affiliate based in South Asia. Initially, Hamid’s recruitment strategy consisted of becoming a regular at neighborhood cafés and conspicuously reading articles or books that he imagined might appeal to a jihadist, in hopes that someone would approach him. “That really didn’t work,” he says. “It was far more effective to be transparent.”

So he started to look for Urdu speakers who seemed like they had time on their hands. When he saw likely candidates chatting with friends on park benches or sipping tea at one of the many outdoor terraces in the Raval district, Hamid would approach them cautiously. “I didn’t want to seem like I was stereotyping an entire population … I think, also, I just didn’t want to get punched in the face.”

He explained that he was a psychologist conducting surveys on people’s strongly held values related to religion, culture and politics. After chatting for a while, he would invite them to take an initial survey designed to assess a person’s level of radicalization, according to three specific criteria: their support of the militant jihadist group Lashkar-e-Taiba; their approval of violence against civilians; and, lastly, their expressed willingness to aid or participate in armed jihad. The survey took 30 to 60 minutes to complete, and Hamid paid everyone who took it €20 ($22) for their time. For the purposes of the study, a person who fit all three criteria was considered radicalized, in which case, Hamid would call them to ask if their friends might also want to take the survey.

As a Pakistani American, Hamid was acutely sensitive to the fact that the people he was approaching might feel profiled. (And in fact, a number of the nonradicalized people who gleaned the thrust of the survey questions were offended, he said.) However, he also recognized the scientific importance of focusing on this particular population.

“We wanted to study radicalization in the context of violent Sunni jihadism, which at the time we conducted our research was the main international terrorist threat,” he explains. It made sense to focus on recruiting from the Pakistani population (and Moroccan population for a follow-up study on the brains of budding radicals) because they represented the two biggest Sunni Muslim groups in the area. And, “the majority of people pulled into terrorist groups from the Barcelona region came from those two ethnic groups,” he says.

The Artis team also believed that it was scientifically important to study groups that weren’t white college students—a population so overly represented in cognitive-science study that they have their own acronym: people from white, educated, industrialized, rich and democratic societies (WEIRD). “Studying sacred values and willingness to fight and die in two separate ethnic groups with very different cultural backgrounds allowed us to examine the generalizability of our claims,” says Hamid.

To protect both the extremists and the study itself, rather than using names, the researchers assigned each volunteer a number. They also tried to avoid asking any questions in the surveys that might put them in tricky legal terrain. “I would tell [the volunteers], ‘Do not tell me anything about a crime you committed, because that will implicate me,’” says Hamid. Instead, the researchers asked hypothetical questions aimed at assessing participants’ beliefs and values, rather than what a person had already done or intended to do with them.

By the end of 2015, Hamid and his team had convinced 146 people to take the survey. He and his colleagues then followed up with the most radicalized of the group—the 45 men who met all three criteria—offering them an additional €100 ($120) to come to a laboratory for the rest of the study. Thirty men, ages 18 to 36, agreed.

Into the Lab

The Autonomous University of Barcelona’s fMRI lab is located in the basement of a blocky gray building flanked by patches of green lawn where, on sunny days, college students like to picnic and read books. There, a team led by Clara Pretus, a neuroscientist in her mid-20s, put these 30 men through the next stages of the study.

The men came to the lab in groups of three or four. After a brief orientation to ease their nerves, the brain scans would begin. The men would lay prone on the bed of the fMRI machine, which would back them into a tube. They wore goggles affixed to a video screen that would flick on and project a statement written in Urdu: “Prophet Muhammad must never be caricatured” or “The Qur’an should never be abused,” for example. Each statement touched on an issue that mattered to the group, based on previous surveys and interviews. The scientists knew which statements aligned with each man’s sacred and nonsacred values, based on those same previous surveys, and they wanted to know how their brains would respond to each. To figure this out, they asked the men to rate how willing they would be, on a scale of 1 to 7, to fight and die for each declaration.

The machine snapped pictures of their brains as the men used a handheld device to make their ratings. After they had gone through all the prompts, Pretus offered them the opportunity to review the slides again—but this time, they’d be able to see how their own responses compared with those supposedly given by their “peers.” This peer group was presented to the men as “the average opinion of the Pakistani community in Barcelona.” But in reality, the researchers had fabricated the ratings for the sake of the experiment. In some instances, the researchers made them appear to align with the men’s responses. In other cases, their “peers” appeared to be more inclined to fight and die for specific values. In still others, less.

After the men had seen how the ratings of their so-called peer group differed from their own, they were given the opportunity to go through the slides one last time—this time outside of the machine—and rate their willingness to fight and die for each statement once again. The scientists wanted to see if the responses from their “peer group” would make them alter their initial responses. In cases where anyone changed his mind, scientists would go back through the fMRI images to see what was happening in his brain as he reviewed the peer information that ultimately compelled him to reconsider his initial answer.

After they completed the final task, the men, whose names they never learned, were free to take their money and go, disappearing into the streets.

Findings

Over the following weeks, the team analyzed the data. As expected, the men expressed greater willingness to fight and die for their sacred values than for their nonsacred values. More interesting were what parts of the brain appeared involved with each question. When participants rated their willingness to sacrifice for their sacred values (defending the Qur’an, for example), parts of the brain linked to deliberation (the dorsolateral prefrontal cortex, inferior frontal gyrus and parietal cortex, which Pretus describes as parts of the fronto-parietal or “executive-control network”) were far less active than when they rated their willingness to kill and die for issues they cared about less (like the availability of halal food in public schools). Dr. Oscar Vilarroya, the lead neuro-scientist on the team, says this indicates that humans don’t deliberate about their sacred values: “We just act on them.”

While this may seem like common sense, the finding was significant, since nearly all sacred-values research to that point had been based on surveys and other tools that assessed what people said—not tied to brain activity. “When you’re taking a social survey, you can lie,” explains Atran. “But brain patterns can’t be faked.” It was the first published study scanning the brains of extremists.

Knowing extremists essentially don’t deliberate when considering the values most important to them confirmed something Atran long believed: that deradicalization programs focused on altering extremists’ beliefs through logic and reasoning, or through trade-offs and material incentives, are doomed to fail. Others had made this argument to explain why programs like France’s civics- and reward-focused deradicalization program, launched in late 2016, had flopped within a year. Here was brain science to support the case.

There was one finding of the study, though, that provided a glimmer of hope for an alternative approach: the areas in the brain linked to deliberation lit up when extremists realized their “peers” weren’t as willing to resort to violence to defend a particular value. And when given the opportunity, post–brain scan, to revise their initial answers to the question “How willing are you to fight and die for this value?” many of them adjusted their rating to better align with their peers. Hamid says this shows that peer groups, like family and friends, play a powerful role in determining whether an extremist will become violent. They will never be able to change the extremist’s core views or values, he says, but they can convince that person that violence is or is not an acceptable way to defend those values. This finding, Atran believes, could have real implications for governments and organizations working in counterterrorism.

“The lesson … is don’t try to undermine their values,” Atran says. “Try to show them there are other ways of committing to their values.”

Critiques and Real-World Applications

The team’s work, published in the Royal Society Open Science journal in June 2019, has garnered a flurry of attention, especially from social psychologists and other academics interested in human motivation. Jonathan Haidt, a social psychologist at New York University and author of the controversial book The Coddling of the American Mind, commended Atran and his colleagues on their “ecological validity”—how relevant the studies are to real-world problems. “We often use the easiest subjects to obtain, which are college students,” he says. “But Scott, at great expense and with great difficulty has always been committed to ecological validity—to studying people who are truly involved in extreme behavior, including terrorist behavior.”

But academics with a background in neuroscience, including Jay Van Bavel, an associate professor of psychology and neural science at New York University, and Patricia Churchland, who studies the intersection of brain activity and philosophy at the University of California, San Diego, expressed more caution. Churchland reviewed the study for Royal Society. In her review, she says, she warned that the brain regions and neural networks from which scientists drew their conclusions are still not very well understood and have been associated with a range of functions beyond simply “deliberation.”

Atran points out that he and his colleagues never set out to map the connection between brain parts and behaviors. Instead, they sought to—and did—find brain patterns that lined up with the results of behavioral studies. (He adds the usual science disclaimer: “All results are tentative, and we look for replication.”)

Meanwhile, as the academic world weighs the research, the Artis team has published additional brain studies on radicalization. And the U.S. military and foreign governments are already plotting how they might put the findings to use. Since the Barcelona work first began, Davis and Atran have been fielding calls from security officials around the world seeking advice on how to deal with radicalized populations and how to apply their research to newer problems, like criminal groups spreading disinformation and taking advantage of weak governance amid the COVID-19 pandemic. Davis is adamant that his researchers steer clear of directly advising any military or government—he doesn’t want the fate of suspects or a nation’s security to be pinned on one of them. But he’s happy to send his colleagues around the world to share their research findings and even collaborate on projects.

And, in a twist, the U.S. Air Force Academy in Colorado got in touch in 2016 seeking to collaborate and study how a cadet’s sacred values and identity with varying groups affect their willingness to fight and die. This April, the Academy, with Artis’ assistance, completed a small study that found that cadets who both viewed religion as a sacred value and strongly identified as a member of a religious group took greater risks than their peers in virtual combat situations. One key takeaway, according to Lieut. Colonel Chad C. Tossell, the director of the school’s Warfighter Effectiveness Research Center, is that the “spiritual strength” of soldiers is as important as the weapons and technology they use. An early draft of the study says the simulation designed for the research could be “useful for selection and training.”

Davis is encouraged by the constant interest he gets from governments, from those in the U.S. to Kenya to Kosovo. The U.S. military continues to aid in funding as the firm sets its sights on the next frontiers: figuring out how and why democratic institutions collapse and how cyberspace is being used to divide people and harden their values, turning nonsacred values into sacred ones. Artis’ work is “first and foremost about field-based scientific research,” and giving policymakers the facts they need to responsibly respond to the problems of the day, Davis says. “We can debate what the meaning of the empirical evidence is, but it’s better to have it than not to have it.”

—With reporting by Mélissa Godin and Madeline Roache/London