Brain learns to anticipate risk, study says

Image credit: Joshua Brown, WUSTL

After the Tsunami in Asia, researchers attempted to explain reports that aboriginal tribesmen had 'sensed' the danger in time to flee to higher ground.

Although some researchers dispute the existence of a "sixth sense" for danger, research from Washington University in St. Louis has found a region of the brain that serves as an "early warning system" -- monitoring environmental cues and helping to change our behavior to avoid danger.

"Our brains are better at picking up subtle warning signs than we previously thought," said Joshua Brown, a research associate in psychology and co-author of a study on these findings in the Feb. 18 issue of the journal Science.

The findings support the idea of a new way to conceptualize the complex processes taking place in an area of the brain called the anterior cingulate cortext (ACC) that is thought to be involved in controlling thoughts and behaviors.

"In the past, we found activity in the ACC when people had to make a difficult decision among mutually exclusive options, or after they made a mistake," Brown said. "But now we find that this brain region can actually learn to recognize when you might make a mistake, even before a difficult decision has to be made. So the ACC appears to act as an early warning system -- it learns to warn us in advance when our behavior might lead to a negative outcome, so that we can be more careful and avoid making a mistake."

The ACC has been studied in great detail in recent years as it is thought to play a critical role in complex and difficult cognitive tasks. Abnormalities in this area are related to numerous mental problems such as and obsessive-compulsive disorder.

The study helps in the development of more effective treatments for such illnesses by providing a clearer picture of the cognitive processes used to monitor and control our behavior.

"Our results suggest how impairment of the ACC mechanisms in schizophrenia can lead to breakdowns in the early warning system, so that the brain fails to pre-empt or control inappropriate behavior," Brown said. "On the other hand, in individuals with obsessive-compulsive disorder, the ACC might warn of an impending problem even when no problem is imminent."

"Interestingly, we also found evidence that the same neurotransmitter involved in drug addiction and Parkinson's disease, namely dopamine, seems to play a key role in training the ACC to recognize when to send the early warning signal," he added.

Additionally, the ACC is thought to mediate between fact-based reasoning and emotional responses lik fear or love.

"For a long time we've been interested in how the brain figures out how to integrate cognitive information about the world with our emotions, how we feel about something," Brown said. "For many reasons, people think the ACC might be the brain structure responsible for converging these different signals. It seems to be an area that's involved in deciding what information gets prioritized in the decision-making process. It seems able to link motivational and affect information – things like goodness or badness – and to use this information to bring about changes in cognition, to alter how we think about things."

Although there is growing agreement that the ACC plays a role in complex cognitions and feelings, several competing theories seek to explain the underlying mechanisms.

Recent research has shown spikes of activity in the ACC at the time when people realize they've made a mistake -- an "oops" moment, or "oh shit!" response. Theories regarding these findings suggest the ACC is primarily involved in detecting and correcting mistakes or conflict.

Brown and co-author Todd Braver conclude, however, that their study provides evidence that the ACC is more accurately thought of as a pre-emptive early warning system that actively works to prevent mistakes.

"We started with the premise that perhaps the cingulate was not responding to the detection of an error or state of conflict, but maybe instead what the cingulate is detecting is the likelihood of making an error," Brown said. "We wanted to see if the cingulate would become more active even in situations where no conflict is presented and no errors are made, but the potential for error is still higher than normal"

Study participants viewed pictures of cars and indicated the direction of arrows by pushing a key. Brain imaging suggested participants learned to associate blue cues with a higher possibility for error, thus providing an early warning signal.

Participants in the study were shown either a white or blue dash which then changed into an arrow pointing to the right or left. They were instructed to indicate the direction of the arrow by pressing one key or another on a keyboard. Conflict was introduced by the occasional inclusion of a larger second arrow that made participants push the opposite button.

"The idea is that at some point you have these competing tendencies – to push the right or left button -- and both are active in brain at same time, which creates conflict," explains Brown. "Some theories suggest that whenever you see these two arrows, then that drives this state of conflict and it's the state of conflict that is being detected by the cingulate."

The researchers varied the delay before the onset of the larger arrow so they could measure how a person would react when they reached the "point of no return" and therefore unable to avoid pressing the wrong key. They also varied it such that for blue lines the participants would respond incorrectly roughtly 50 percent of the time, whereas for white lines the error rate was around 4 percent.

The researchers used functional magnetic resonance imaging (fMRI) to capture images of brain activity at 2.5 second intervals.

"We didn't tell them that the white or blue cue offered any clue about their likelihood of making an error on any particular trial, but by the end of the session, some of them had begun to figure it out, at least on a subconscious level," Brown said.

Even participants who remained relatively unaware of the significance of the blue cue showed increased activity in the ACC in response to the blue color.

"It appears that this area of the brain is somehow figuring out things without you necessarily having to be consciously aware of it," Brown said. "It makes sense that this mechanism exists because there are plenty of situations in our everyday lives that require the brain to monitor subtle changes in our environment and adjust our behavior, even in cases where we may not be necessarily aware of the conditions that prompted the adjustment. In some cases, the brain's ability to monitor subtle environmental changes and make adjustments may actually be even more robust if it takes place on a subconscious level."

The study is also significant because of its use of computer models to accurately predict the patterns of brain activity observed in the experiment.

"We started by building a detailed computer simulation of the ACC, and then we found that the computer predicted the existence of the early warning signal in ACC," Brown said. "This was an exciting result, but we still needed to test the prediction in humans to demonstrate that the model prediction was correct."

They also used a second computer model that had been previously developed to support another theory of the ACC.

"By simulating both models we could then adjudicate between them and do so in a way where we forced each one to make predictions that we only tested after the fact," Brown said. "By integrating the theory, the computational simulation and then the fMRI testing, we are providing other scientists with some very rigorous evidence that our new theory is accurate."

The use of omputer modeling as a means to understand the brain has increased over the past two decades, said Brown.

"In fact, our computer model also makes some other exciting predictions about how the ACC works, but we haven't had an opportunity to test them yet," Brown said. "We've got our work cut out for us."

Related Stories