The brain’s reward system learns the actions that produce positive outcomes, such as obtaining food or sex. It then reinforces the desire to initiate those behaviors by inducing pleasure in anticipation of the relevant action. But in some circumstances this system can become oversensitized to pleasurable but harmful behaviors, producing pathological impulses like drug addiction, binge eating and compulsive gambling.

But what if we could spot impulsive urges in the brain and intervene to prevent the act? This is the promise of a new study published December 18 in Proceedings of the National Academy of Sciences, led by neurosurgeon Casey Halpern, of Stanford University. His team identified a “signature” of impulsive urges in part of the brain’s reward-learning circuitry, the nucleus accumbens. Delivering electrical pulses to this region on detecting this activity reduced binge-eating behavior in mice. They also observed the same signature in a human brain, suggesting the technique has potential for treating a range of conditions involving compulsive behaviors. “We’ve identified a brain biomarker of loss of control,” Halpern says. “If we can use that to prevent any of these potentially dangerous actions, we can help a lot of people.”

Researchers used a variation on deep-brain stimulation (DBS) in their experiments, a well-established treatment to diminish the shaking present in Parkinson’s disease that is also showing promise in other conditions including depression and obsessive-compulsive disorder. Exactly how DBS has beneficial effects is still being debated, but there can be side effects. When treating movement disorders, patients may experience tingling and muscle contraction, says neurosurgeon Tipu Aziz of the University of Oxford. The long-term consequences in other regions are unknown but could include seizures, or effects on cognition, he says.

Ordinarily, DBS stimulates a brain area with an uninterrupted current. But researchers are investigating ways to provide electrical pulses only when needed by monitoring telltale brain activity related to a particular condition or symptom. This technique, known as “responsive neurostimulation” (RNS), has proved effective for epilepsy and an existing system has been approved by the U.S. Food and Drug Administration for treating partial onset seizures (which affect only part of the brain). There is also some evidence RNS may be better than continuous stimulation for treating Parkinson’s, Aziz says.

Halpern’s group had previously shown standard DBS reduces binge eating in mice by as much as 50 percent. To provide as-needed stimulation, they set out to find a brain “signature” for impulsive urges they could use to trigger stimulation. They gave six mice high-fat food pellets one hour a day for 10 days (after which all the mice were binge eating). They recorded activity from the animals’ nuclei accumbens before and after this “learning” period and when they ate their normal diets. They saw an increase in low-frequency (“delta band”) brain-wave activity in the nucleus accumbens, which peaked one second before a mouse indulged in a binge session. This activity did not appear at the start of the experiment or when the animals were eating standard chow that was not high fat. They also demonstrated this signature was specific to binge eating by showing it was not seen immediately before interacting with a young mouse, which mice also find pleasurable. “We don’t want to block natural rewards,” Halpern says. “We’re not going after anybody with an impulse—I like a glass of wine myself.”

Having identified a “moment of weakness” signature, the team next tested whether using it to trigger electrical simulation to disrupt the impulse reduced binge eating. They compared automatically triggered stimulation with continuous, random, and manual stimulation in which a researcher activated the electrode on seeing the mouse move to start eating. All except random stimulation reduced the amount consumed. But they also showed only continuous stimulation decreased the amount of time mice spent interacting with young mice, considered to be an undesirable side effect.

The team next turned to a human brain to see if the signature they had discovered in mice also applied to people. They were able to conduct their search in an obsessive-compulsive disorder patient who was unresponsive to other treatments and had opted to undergo surgery for DBS. They monitored an electrode implanted in the man’s nucleus accumbens while he performed a task in which he had to push a button when a visual target flashed on a screen to receive a cash reward. Once he was accustomed to the task—and the reward from receiving cash—they saw a similar increase in delta activity to the one observed in mice, brain waves that intensified immediately before he began a task.

The fact similar brain activity was seen in both species for behaviors aimed at different rewards suggests this signature may be common to many compulsive behaviors. Of course, nobody with a gambling habit would undergo neurosurgery nor would physicians consider it. “We’re going after people who are dying of their condition or are severely debilitated,” Halpern says. “People who are about to have gastric surgery or those that fail it because they can’t stop binge eating, or alcoholics.”

One potential problem is if the signature of a craving is too specific, it may not be relevant to multiple disorders or even be applicable outside the lab for routine clinical use. “It’s not entirely clear whether this signal is related to the [experiment/task] or the behavior, “says psychiatrist Damiaan Denys of the University of Amsterdam, who was not involved in the study. If it’s too general, you could imagine a situation where someone being treated for compulsive eating can no longer enjoy playing blackjack. “But it’s important because it shows it’s possible to detect a signal that may be used for psychiatric disorders,” Denys says.

This is preliminary research, involving only one human, without showing a human treatment effect. “It was a first-in-man case study to validate our mouse studies—now we have to do it in a larger sample,” Halpern says. “We have an application with the NIH [National Institutes of Health] to test this in obese patients that exhibit bingelike behavior.”

This article was originally published at Scientific American.