Teenagers on June 8 visit the site where three Clarksburg High School students were killed in a crash the night before. The CDC says the risk for accidents increases with teen passengers in the car. (Amanda Voisard/For the Washington Post)

More than a decade ago, the National Institutes of Health sent researchers to high schools across suburban Washington to track teens with lasers and video cameras.

They wanted to see how their young subjects drove when they had passengers with them.

So the researchers stood outside 10 parking lots in Maryland and Virginia. The teens had no idea they were being watched as researchers jotted down details about who was riding in the cars and then monitored the subjects as some sped and tailgated away from campus.

Now, expanding on that work, scientists are using driving simulations and brain scans to try to explain not just how young people drive when they’re with others, but why.

“We don’t want teenagers just staying at home. We want people to go out and explore and figure out who they are,” said Emily Falk, who as director of the Communication Neuroscience Lab at the University of Pennsylvania has been putting teen drivers in MRIs to probe their minds.

But the cognitive upsides of youth can also be a vulnerability. As teens, “we’re really sensitive to the social environment, and that’s mostly a good thing,” Falk said. But “this kind of sensitivity to the social environment can lead to risk-taking.”

Despite decades of progress, traffic accidents remain the No. 1 cause of death for young people, killing 2,600 teens in 2014, according to the Centers for Disease Control and Prevention.

Although fatalities in crashes involving young drivers fell by nearly half from 2005 to 2014, federal safety officials on Friday said that figure jumped 10 percent in 2015, helping spur a sharp increase in traffic deaths overall.

On the Tuesday before graduation last month in Montgomery County, one of the original counties studied by the NIH back in 2004, three teammates on the Clarksburg High School football team were killed when their Ford F-250 drove off a dark, curving road and into a tree.

“The risk of a crash is higher when there are teen passengers in the car,” said Ruth Shults, a CDC epidemiologist and teen-driving expert. “The risk is there, but it’s not there all the time.”

Generally, teen passengers are not a distraction, Shults said, so it’s important “we don’t get the idea that, ‘Oh my gosh, every time we put two kids together it’s going to be disaster.’ ”

Problems arise when risks are layered atop one another, like driving with passengers and at night, a combination that can be treacherous for inexperienced drivers, she said.

Success in reducing teen deaths has largely come from paring back opportunities for disaster. Safer cars, plus the widespread adoption of graduated driver-licensing programs, which can delay driving with and without passengers and impose some restrictions on nighttime travel, have made a big difference. Tightening those standards in states with looser rules would save many lives, experts said.

Getting a closer look inside teen cars and brains can also hone prevention efforts, according to researchers.

The NIH snooping outside the Montgomery and Fairfax County high schools netted intriguing results and marked the beginning of a snowballing research effort. The tracking was legal, but researchers got permission from principals anyway.

“You get this nice pattern,” said Bruce Simons-Morton, who helped lead the study: Teens driving male passengers were more than three times as likely to speed as those who drove alone or with female riders.

It made the researcher want to better understand what’s happening between young drivers and their passengers. “Pretty much, the research usually leads to new questions,” said Simons-Morton, who was a professor in Texas studying high-school drinking before moving to Bethesda and joining the National Institute of Child Health and Human Development in 1992.

So he partnered with colleagues at the University of Michigan Transportation Research Institute, and they recruited Ann Arbor-area teens for another somewhat-stealthy study.

This time, they brought in a group of 16- to 18-year-old boys who recently got their driver’s licenses and told each one that another high-schooler would join them for a driving simulation.

What they got instead was a male college student who was in on the ruse and equipped with a script.

“What we’re trying to get at here is: Under what conditions do passengers increase or decrease risk among teen drivers?” Simons-Morton said.

The young-looking actors — researchers prefer to call them “confederates” — played the part of either “risk-averse peer” or “risk-accepting peer.”

One would make a tardy arrival and say, “Sorry I was a little late getting here. I tend to drive slowly, plus I hit every yellow light.”

The other, also tardy, would say, “Sorry I was a little late getting here. Normally I drive way faster, but I hit like every red light.”

They’d watch videos together of either safe or aggressive driving, and the actor would make it clear he preferred one or the other, depending on the role he was playing.

Finally, with the subject sitting in the passenger seat, the actor would take the wheel on the simulator, either stepping on the gas and racing through yellow lights or accelerating gently and stopping ahead of the yellows.

Then the teen guinea pig was told it was his turn to drive.

The results?

The teenagers who had been hanging out with the aggressive actors ran more red lights and made other irresponsible maneuvers. “They drove in a much more risky way,” Simons-Morton said.

In a similar experiment with another batch of teens, overt peer pressure appeared to push young drivers to act in riskier ways. That’s a common finding.

“If somebody encourages you to buy this rather than that, encourages you to have another drink, almost anything, it has an influence on us,” Simons-Morton said.

But this was different. In this study, the subjects weren’t being pressured or persuaded by peers to drive faster. They were simply being primed ahead of time. The teens appeared to be swayed by their perceptions of other people’s attitudes, or what Simons-Morton calls “social norms.”

It hinted at an old parental trope. “You want them to be with good friends, not with the . . . risky kid,” he said.

Still, there was more to learn about how teen brains trip their way toward dangerous driving, which brought Simons-Morton to another flavor of the same experiment.

One of the activities lined up for the same Ann Arbor subjects, a week before they met their speeding or unhurried pretend peers, was a stop in Falk’s lab for an MRI.

The teens had their brains gauged while they played Cyberball, a ball-tossing video game social scientists can use to stir up feelings of isolation. The player believes he’s in a fair game, but the computerized partners eventually stop throwing him the ball.

“It makes people feel bad,” Falk said.

Meanwhile, Falk was tracking areas of the brain known to process “social pain.”

“Young boys in particular don’t want to tell you they got upset they were excluded. But their brain doesn’t lie,” Simons-Morton said.

Using that information, the researchers could predict which boys were likely to drive in risky ways after interacting with the more aggressive driver in the simulation.

“The kids who are the most sensitive in the brain scanner when they’re being excluded are then also the kids that go on to take more risks when they are driving with a passenger,” Falk said. “They’re speeding up to gun it through the yellow light.”

Falk said she understands their predicament. If it’s “particularly costly for you to be left out . . . it may not actually be that crazy for you to take risks to try to fit in.”

She and others want to flip that phenomenon on its head to make safety campaigns more effective. Already, many teens are a good influence in the car, as elsewhere. And that can be tapped further.

Falk said another twist on the same Ann Arbor simulation experiment shows a link between brain activity involved with self-control and some teens’ approach behind the wheel. “Those kids . . . drove in safer ways when they were with peers who support that,” Falk said. “Peers can serve positive purposes.”