The brain mechanisms for predicting future events that help us make our choices have been dissected. All of this takes place in the anterior cingulate cortex.
- Our decision-making is based either on a model, where we imagine the consequences of the action, or for lack of a model, on the experience of past actions.
- Observing the anterior cingulate cortex, an area of the brain involved in decision-making, allows us to predict our future decision
Before making a decision, our brain begins by imagining what may happen. For example, before leaving for a restaurant, he will imagine the route to take to get there as quickly as possible. For the first time, researchers have revealed the sophisticated mental machinery that helps the brain simulate the results of different actions and make the best choice. They published the results of their research on November 4 in the journal Neuron.
Two brain decision-making mechanisms
The researchers found that the anterior cingulate cortex, involved in decision-making, is also the one that visualizes the consequences of a decision to help us make the best possible choice. “We managed to identify a brain structure involved in learning and demonstrate that its activity encodes multiple aspects of the decision-making process.”, added Thomas Akam, professor at the University of Oxford and lead author of the article. “These results are very excitingcontinues Rui Costa, another author of the study. These data identify the anterior cingulate cortex as a key brain region in pattern-based decision-making, more specifically in predicting what will happen in the world if we choose to do a particular action over another..”
Not all decisions come under this cerebral mechanism, say the researchers, and some result from learning without a model, therefore without simulation of the consequences of the action, which is based on the experience of past actions. When a route, for example, is known, our brain will use model-free learning and follow the route usually taken.
The choice reveals the learning strategy
To isolate these two cognitive patterns, model-based and model-free, the researchers set up a two-step puzzle for the mice. The animal begins by choosing one of the holes in the center to stick its nose through, which activates one of the other two holes. Each of these holes has a probability of providing a glass of water. “Just as in real life, the subject must perform long sequences of actions, with uncertain consequences, in order to obtain the desired results.”, specified Thomas Akam. The mice had to seek to understand two key variables: which hole is most likely to provide a glass of water, and which hole activates the hole that provides the glass of water. Once the task was learned, the mice chose the action sequence that offered the best result. The researchers then modified the experiment so that the animals were flexible. Occasionally, the side port most likely to deliver a drink has flipped or the correspondence between the center and side ports has reversed.
The mice’s choices during the study revealed their learning strategies. “Model-free and model-based learning should generate different choice patternsdescribed Thomas Akam. By looking at the behavior of the subjects, we were able to assess the contribution of either approach.” The study of the results, which was based on the analysis of approximately 230,000 individual decisions, showed that the mice use the two approaches in parallel. “This confirmed that the task is suitable for investigating the neural basis of these mechanisms.continued Dr. Costa. We then moved on to the next step: studying the neural basis of this behavior.”
Turning off the anterior cingulate cortex unseated the mice
The researchers focused their attention on the anterior cingulate cortex, the region of the brain where decision making is constructed. “Previous studies have established that this region is involved in stock selection and provided evidence that it may be involved in model-based predictions.said Rui Costa. But no one has checked the activity of individual neurons in this region in a task designed to differentiate between these different types of learning..”
The researchers found a strong link between the activity of these neurons and the behavior of the mice. By observing patterns of activity between groups of cells, the scientists were able to decode whether the mouse is about to choose one hole or another. “This provided direct evidence that the anterior cingulate cortex is involved in making predictions based on patterns of the specific consequences of actions, not just whether they are right or wrong.”, welcomed Dr. Akam. Turning off these neurons in mice prevented them from reacting flexibly as the situation changed.
.
