In a recent study, researchers at the Massachusetts Institute of Technology (MIT) have identified a specific brain circuit that helps people learn through the experience of others.

Observational learning is a form of learning that occurs through the observation of other people. It acts as an important evolutionary advantage, allowing people to avoid potential danger without having to experience it directly.

“So much of what we learn day-to-day is through observation,” Kay Tye, associate professor of brain and cognitive sciences at MIT and a member of MIT’s Picower Institute for Learning and Memory, said in a statement. “Especially for something that is going to potentially hurt or kill you, you could imagine that the cost of learning it firsthand is very high. The ability to learn it through observation is extremely adaptive, and gives a major advantage for survival.”

Interested in how this valuable form of learning occurs, a team of researchers led by Tye conducted a study to identify the specific brain functions that enable observational learning.

Their research is published in the journal Cell.

Through their research, the team was able to distinguish a brain circuit between the parts of the brain known as the anterior cingulate cortex (ACC) and the basolateral amygdala (BLA) that plays a pivotal role in observational learning.

Notably, these two regions of the brain are both active in social interaction more generally. The ACC is involved in social evaluation, while the BLA is associated with processing emotions.

“We knew from previous research that both the ACC as well as the amygdala are important for observational learning in rodents as well as humans,” said Stephen Allsop, a former M.D./Ph.D. student at MIT who co-led the study. “However, it is not known how neurons in those regions actually encode information during observational learning.”

The brain circuit

In the study, the researchers evaluated the function of the brains of mice, who share many similarities in brain structure and connectivity with humans, during social learning.

They observed the brain activity of mice as they witnessed another mouse being shocked. The shock was accompanied by a cue, such as a tone or a light, that signaled the danger. The mice being shocked come to associate the shock with the cue, and will freeze when they hear or see it.

One day later, the researchers noticed a similar reaction to the cue in the mice who had witnessed other mice being shocked, but were not shocked themselves. Just like those who were directly shocked, they too were conditioned to fear the cue. Essentially, the researchers were watching the process of observational learning play out before them.

In order to understand how the process of observational learning occurred in the brains of the mice, they recorded electrical activity in the ACC and the BLA of the observer mice as they learned to associate the cue with danger.

Then, using a new type of analysis called a neural trajectory analysis, they analyzed the firing rates of the neurons during the learning process. This analysis revealed a high level of activity in the ACC as a mouse observes another mouse being shocked. Afterwards, information is sent from the ACC to the BLA, where it is processed, associating the cue to the shock.

This demonstrated an important connection between the ACC and BLA that is fundamental in the process of observational learning.

“Through our experiments we were able to record from circuit-specific neurons and show that neurons in both regions encode the predictive cue as well as the distress of the demonstrator,” said Allsop. “We also showed for the first time that a subset of amygdala neurons actually require input from the ACC neurons during learning in order to encode the predictive information about the cue that drives learning.”

Testing the connection

Once the researchers identified the neurons that connect the ACC to the BLA, they blocked the connections between these neurons during the observational learning task. To do so, they used optogenetics to express an inhibitory protein that stopped the neural activity between the ACC and the BLA. When the connections were blocked in mice, the mice did not learn to fear the cue.

Blocking this connection in mice receiving the shocks accompanied by a tone, however, did not have any effect on the observer mouse’s ability to connect the cue and the shock. This suggests that the ACC-BLA connection is specifically associated with socially-derived information.

“Not only did amygdala neurons stop encoding the cue, but behaviorally, animals also did not learn,” said Allsop. “Interestingly, this circuit did not impact learning when animals learned from their own experience.”

The researchers also found that the ACC-BLA connection was active in other types of social learning and behavior, including learning how to interact with other mice. The connection is active, for example, when mice learn to fear an aggressive mouse after observing it interact with other mice.

What’s next?

Moving forward, Tye’s team intends to research the role of this connection in learning rewards rather than fear.

“Observational learning is still a relatively new field,” said Allsop. “Most of the work has focused on observational fear learning but it has been more difficult in certain animal models to get detailed circuit level information about observational reward learning functions. This is one potential direction the Tye lab is interested in pursuing.”

Allsop suggested that understanding how these neural connections function could also provide insight into certain psychiatric diseases associated with abnormal social behavior.

“The burden of social deficits seen in psychiatric diseases such as autism and social anxiety disorder provide motivation to understand how abnormal neural function gives rise to aberrant social behavior,” he said. “Another potential way forward from this project is to use the information we have now learned about how this circuit functions at baseline to ask questions about how this circuit might function differently in various genetic mouse models of psychiatric diseases with social symptoms such as autism and social anxiety disorder.”