In a secluded section of Raleigh, North Carolina live 16 carefully selected recruits for a high-priority Pentagon project that could play a life or death role in Afghanistan. 

None of them have any previous military experience, so their training is necessarily disciplined – if they get through the programme their work could pave the way for others, like them, who will be called on to make life-saving decisions in high-stress environments. As a result, nothing is left to chance.  Their days involve intense physical and mental training to see if they are able to cope with the rigors of war.

Outside, a trainer introduced me to one recruit, Jimmy, who is outfitted with a telemetry vest, which can measure his physiological changes, such as heart rate, breathing, and skin temperature. It gives him an almost Olympian look, like an athlete whose every breath is measured by a team of coaches.

But you also get a sense that Jimmy gets a kick out of it. As he is ordered forward, he is panting and his tail is wagging with the enthusiasm of youth.

The good-natured black Labrador retriever is the Jason Bourne of dogs – part of a specialised Navy programme designed to select and train the most highly skilled members of his breed for a high-priority mission: finding improvised explosive devices (IEDs) on the battlefield. While prototypes of the dogs have already been deployed, research on Jimmy and his colleagues is part of a “2.0” version of the programmme, which is based on formal science studies being conducted here at North Carolina State University’s College of Veterinary Medicine, and other universities.

Over the past six years, the Pentagon has spent over $18 billion trying to find ways to detect and neutralise roadside bombs – the leading killer of US troops in Afghanistan –by investing in everything from better mine detectors to newfangled devices that shoot beams of energy. But in North Carolina, a state that is home to a 300-year hunting dog tradition, everyone seems to agree on the best technology for the job.

In the lab, I’m introduced to Dakota, one of Jimmy’s colleagues. In front of Dakota are two small containers, one scented with vanillin and the other with ethanol. Dakota, who is learning to signal when she detects a scent, is supposed to select the vanillin, a training aid that will later be switched to ammonium nitrate used in explosives. If she selects correctly, she’ll be rewarded with a treat. She gets the first one right, choosing vanillin, misses the next two, but then got the fourth one right. At that point, she was averaging 50-50, which is not good, but she was still a beginner with time to learn.

Biting machine?

Though war dogs may be high profile these days – thanks in part to Cairo, the Belgian Malinois who accompanied Navy Seals on the Osama bin Laden raid – there were virtually no deployed military bomb dogs working on battlefields when the US invaded Iraq, and hadn’t been for decades. After the Vietnam War, the military’s working dog programme was handed over to the Air Force, which used them to patrol missile fields in the United States. Those dogs were used to being guided on leashes, living in air-conditioned kennels and working limited shifts: a far cry from a war zone.

That quickly changed with the arrival of IEDs and in 2007, 13 dogs – all Labs – were deployed with the Marines to Iraq and Afghanistan. When the dogs started finding bombs, Marines asked for more, and today there are some 600 dogs.

Whilst successful, they were all essentially prototypes for the next generation of bomb dog. “After we do all this science, we’re going to rewrite the protocols,” says Lisa Albuquerque, an energetic Navy official who heads the programme, known formally as the IED Detector Dog. “It’s basically an upgrade to an existing model.”

Her starting point was talking to the Marines who would ultimately use the dogs. She recalls asking a staff sergeant: "Do you want a dog that bites," to which he replied: "I have 13 heavily armed Marines, why do I need a dog that bites?" What they wanted, it turned out was a good-natured dog, hence the Labradors. From there, she began an extensive array of experiments with researchers at North Carolina State University, Duke University, and Oklahoma State, to develop the training and selection criteria that will make them the perfect IED detector.

Talking about dogs as a technology, like a rifle or a pair of binoculars, is part of what makes the Navy programme unique. “I think that a real key point is that the science and technology that the [Department of Defense] has funded [in the past] has never been about trying to optimise the dog,” says Albuquerque, who spent 25 years as a military dog trainer. “It’s all been about trying to make a sensor that’s as good as a dog, and so our approach, which is really very unique, is saying, ‘Wait a minute, for the near term, dogs are still going to be the best sensor.’”

Mind games

But smell is just one part of the Navy’s dog programme – a large part of the effort is animal cognition. Brian Hare, who heads Duke University’s Canine Cognition Centre, is interested in animals’ inferential abilities, or the ability to reason and problem solve. Until recently, it was something that scientists thought was impossible.

Hare demonstrates his work with a video, showing Yaya, a great ape in Uganda, who is presented with a mental puzzle: a long plastic tube that is latched down and can’t be moved. “At the bottom is the equivalent of a Peabody award,” says Hare. “It’s three peanuts.”

The question is how would Yaya get the peanuts (and could I come up with a solution before she does)? In less than 10 seconds, before I could even venture an answer, Yaya has already solved the problem. She goes to her water dish, sucks some water into her mouth, and then spits the water into the tube, allowing the peanuts to float to the top.  Pretty much all of the female apes, like Yaya, solved the nut problem the same way, according to Hare. “There’s also a male solution to this problem,” he adds. “Many of the male chimpanzees just peed in the tube.”

Hare has extended these cognition tests to dogs. In one test, for example, Border Collies were shown several objects they know by name, such as tennis balls and Frisbees. The dogs then were told to fetch the “gazzer”, a made up word for a toy car, the one object whose name they don’t know. The dogs consistently fetched the toy car, which Hare attributes to the dogs reasoning that the “gazzer” must be the one object they didn’t know: an example of inferential reasoning. “It’s even scarier than that because there’s evidence that dogs, after only two pairings of hearing the word out loud – no food rewards, no training, no nothing – remember the words for as long as four months,” says Hare. “That’s on the order of what human children do.”

Just understanding that dogs are smart isn’t enough, however. At North Carolina State University, Barbara Sherman, a specialist in animal behaviour, and David Dorman, a veterinary toxicologist, are focusing on coming up with tests that can predict which dogs will be the best bomb dogs, in the way that the SATs help predict which students will do well at college. Good attributes of a working dog might be factors such as intelligence and motivation. “Some of these factors we could try to measure directly, some we’re not able to measure directly,” says Dorman, who is wearing a red tie adorned with dogs. “It’s very similar to our student situation.” 

The tests include seemingly simple tasks, such as how a dog reacts to climbing industrial stairs (many dogs don’t like staircases you can see through to the ground), or measuring their reaction to gunshots. The researchers will also soon be testing the dogs in a new lab equipped with cameras, speakers that pipe in white noise, and odour sources. The room, which includes obstacles for the dogs to work around, will test to see which dogs work best on their own, or “off leash”, and which can create a good spatial map – not something that has traditionally been looked for or rewarded in working military dogs – but could prove critical for bomb dogs.

But as researchers begin to understand more about the canine mind, there are also signs that, like humans, they are susceptible to conditions like Post-Traumatic Stress Disorder. That means that on top of all of the intelligence tests, physical training and scent trials, researchers must also try to identify dogs that are resilient, even when faced with prolonged exposure to explosions and gunfire. For example, in one test, to measure “resiliency” in a war zone, the researchers study the dogs’ reaction to the sound of a gunshot.  Why is this important? “You might be less motivated to be an enthusiastic participant in hunting for olfactants if you’re afraid you’re going to be blown up,” says Sherman.

If all goes well, the first 2.0 detector dogs will be deployed in Afghanistan early next year.  The ultimate question, however, is whether these new and improved bomb dogs are more effective. Albuquerque says that for now, the best proof lies with the troops: When Albuquerque first spoke with Marines at the beginning of the programme, they were reluctant to even consider dogs. Their attitude quickly changed when the dogs started finding IEDs, and in a show of respect, early dogs in the programmme, like Cann, were named after fallen Marines (Cann was named for Staff Sergeant Adam Cann, a dog handler killed in Iraq; Cann, the dog, is now retired and living with a former Marine).

Marines, who once doubted the utility of the dogs, also now volunteer to carry provisions for the canines. “They’re not carrying water because they like the dog; they’ll leave the dog behind,” says Albuquerque. “If they’re carrying extra water, they believe in it and it works.”

If you would like to comment on this article or anything else you have seen on Future, head over to our Facebook page or message us on Twitter.