Dean Mobbs, a neuroscientist at the Medical Research Council in Cambridge, England, wondered if humans have similarly layered fear responses. He and his colleagues were not about to send people into tiger-infested meadows, so they designed a clever alternative: They programmed a survival-themed video game that subjects could play while lying in an fMRI scanner. The game is similar to Pac-Man. You see yourself as a triangle in a maze and press keys to maneuver through it. At some point a circle appears. This is a virtual predator being guided by an artificial intelligence program to seek you out. If the predator captures you, you receive a small electric shock on the back of your hand.

This deceptively minimalist predator-prey game triggers some remarkably intense feelings. Mobbs measured the skin conductance of his players by rigging them up to a device similar to a lie detector. He found that when the predator was bearing down on players, they often experienced the same changes to their skin as those seen in people having panic attacks. Mobbs unleashed two kinds of predators on his players, a less adept one that was easy to escape, and a smarter one that was more likely to capture its victim. When people were chased by the better predator, they showed a stronger panic response in their skin, and they also crashed into the walls of the maze more often.

Meanwhile, striking changes were happening inside the brains of the players. The predators would first appear on the far side of the maze. While they remained at a distance, the same brain regions tended to become active in the players, a network that included parts of the amygdala as well as some other structures in the front of the brain. But when the predator was closing in, those brain regions shut down and a network of previously quiet regions farther back in the midbrain became active.

Mobbs’s results mesh nicely not only with the work of the Blanchards but also with some other, more recent studies of rat neurology. For example, one of the midbrain regions that Mobbs and his colleagues observed becoming active in humans when a “predator” was close is an area called the periaqueductal gray region. This area showed higher activity in the people who crashed into the walls more often, providing further evidence that it plays an important role in panic. Researchers have explored the anatomy of fear more directly in rats; by manipulating different areas of the rat brain, they are able to alter parts of the standard fear-driven sequence of behavior. When neuroscientists put electrodes into the periaqueductal gray region of rat brains and stimulated the neurons there, the creatures immediately started to run and jump uncontrollably.

Fear, the new results suggest, is not a single thing after all. Rather, it is a complex, ever-changing strategy mammal brains deploy in order to cope with danger. When a predator is off in the distance, its prey—whether rat or human—powers up a forebrain network. The network primes the body, raising the heartbeat and preparing it for fast action. At the same time, the forebrain network sharpens the brain’s attention to the outside world, evaluating threats, monitoring subtle changes, and running through possible responses. Another important job it performs is keeping the midbrain network shut down so that, instead of fleeing at top speed, a prey animal keeps very still at first. As the predator gets closer, however, the forebrain’s grip on the midbrain loosens. Now the midbrain becomes active, orchestrating a powerful, quick response: fight or flight. At the same time it shuts down the slower, more deliberative forebrain. This is no time for thinking.

It may be unsettling to find that our brains work so much like a rat’s. But the amygdala and the periaqueductal gray are ancient parts of the brain, dating back hundreds of millions of years. Our small hominid ancestors probably faced the same kinds of threats that baboons do today from leopards, eagles, and other predators. Even after we evolved the ability to use weapons and became predators ourselves, this ancient brain circuit still offered a useful defense against members of our own species.

Unfortunately, our exquisitely sophisticated brains may make this predator-defense circuit vulnerable to misfiring. Instead of monitoring just the threats right in front of us, we can also imagine threats that do not exist. Feeding this imagination into the early-warning system may lead to crippling chronic anxiety. In other cases, people may not be able to keep their periaqueductal gray and other midbrain regions under control. As we perceive predators getting closer, our brains normally make the switch from the forebrain to the midbrain regions. People who suffer panic disorders may misjudge threats, seeing them as far more imminent than they really are.

To test these possibilities, Mobbs and his colleagues are beginning to study people who suffer from fear-related disorders as they play the predator game. Such work may not uncover a biological distinction between angst and the heebie-jeebies, but it may show how much better we can understand ourselves—and tame our inner demons—once we appreciate the many dimensions of fear.