Article summary of Adolescent-specific patterns of behavior and neural activity during social reinforcement learning by Jones et al. - 2014 - Chapter

What is this article about?

Humans need to be able to decipher and learn from social signals. Especially during adolescence, the impact of social signals is magnified. When pre-adolescents enter adolescence, they tend to spend more time with their peers, and less time with their parents. They also rely more on their peers than on their parents for guidance and approval. They also show increased attention and higher neural activation in response to peer acceptance. When adolescents feel related to others and when they feel accepted, this is associated with higher self-esteem, and better adjustment in school. In contrast, when adolescents feel rejected, this can lead to school withdrawal, aggression and mental health problems.

Peers impact an adolescent’s behaviour in different ways. For instance, when there are peers in the car, accidents increase. This is only true for adolescents. When there are peers around during a decision-making task, this also leads to higher risky decision-making in adolescents compared to children and adults. Furthermore, when adolescents feel rejected by their peers, they are more likely to engage in risky behaviors, to fit in with the group. One explanation / theory is that feedback from peers serves as a reinforcer, a reward. Research has shown that the neural circuitry that evaluates social (praise, gain, positive affect) rewards overlaps with a neural circuitry that evaluates nonsocial (money, food) rewards. In adults, it has been shown that the ventral striatum (the reward area in the brain) supports learning from social feedback from peers.

In the current article, the goal is to evaluate differences across different age groups in social reinforcement learning from peers, and 8 to 25 year olds will participate. The goal is to determine whether adolescents, compared to children and adults, learn to associate different peers with distinct probabilities of receiving positive feedback. This will be measured with fMRI.

Which methods were used?

Experiment cover story

The experiment consisted of two sessions. In the first session, the participants were lead to believe that they would receive social feedback from peers during a task that they would complete on the second visit. They were shown 5 photographs of gender-, age-, and ethnicity-matched peers. They selected three peers that they would like to interact with. They rated the peers on a scale from 0 to 10 for likeability and attractiveness. Furthermore, participants completed a personal survey where they reported information about themselves. Then, the participants were told that each of the three selected peers would see their survey and the surveys of two others. These peers would write notes on their surveys. Each note that the peers would write would indicate something positive, because the peers had a limited number of notes to write. Notes from peers thus became ‘rewards’.

In the next session, the participants were told that the notes were collected and they would be shown how often the peers decided to write notes to them. However, the delivery of notes was manipulated. Each of the three peers was associated with a distinct probability of social reinforcement, namely ‘rare interaction’ which was defined by 33% of notes and no positive social reinforcement on 66% of the trials. No positive social reinforcement means that the peers decided to write notes for others, but not for the participants. Further, there could be ‘frequent interaction’, which refers to that there was positive social reinforcement on 66% of the trials and no positive social reinforcement on 33% of the trials, and lastly ‘continuous interaction’, which was defined as 100% positive social reinforcement.

What can be concluded?

This experiment showed that there were adolescent-specific age differences in reinforcement learning behavior and neural response patterns. In children and adults, different amounts of positive social reinforcement enhanced learning. In adolescents, all positive social reinforcement lead to lower positive learning rates and higher activity in their response planning circuitry. This means that adolescents do show a unique sensitivity to peers, but adolescent behavior in social contexts is not only explained by simple reinforcement learning theory.

Positive learning rates showed a quadratic patterns, which can mean two things:

1. Adolescents do not learn to discriminate between cues that are associated with different amounts of positive social feedback, or;
2. Adolescents’ behavior is not captured by simple reinforcement learning predictions.

Adolescents demonstrate lower positive learning rates, but this does not mean that they do not learn. At the end of the experiment, adolescents as well as children and adults all rate peers who gave them positive feedback as more likeable. There was also no difference in observed negative learning rates across age groups. The adolescent positive learning profile could be explained by an overall vigilance to the receipt of peer approval. This means that a close friend, but also an anonymous or unknown peer can enhance adolescents’ risk-taking behavior. Lower positive learning rates could also be explained by increased motivation toward which is the socially least reinforcing, which is in line with work that suggests that adolescents engage in risky behaviors when they perceive themselves to be less socially accepted. It was also shown that the anterior to mid insula response is correlated with positive prediction error fluctuations and that this is higher than in children and adults. The insula is responsible for processing subjective feelings and awareness about one’s body, feelings of distress, and overall processing of affective states that are the result of interacting with other people. This non-linear finding in the insula supports the hypothesis that peer approval is emotionally salient to adolescents. Furthermore, the data showed that adolescents, more than children and adults, activated regions with response planning circuitry when they received positive social approval, regardless of who the peer was. The data also showed that the ventral striatum and medial prefrontal cortex were equally engaged across age during social reinforcement learning. This means that fundamental reinforcement learning mechanisms support social reinforcement learning from late childhood to adulthood. It seems that the heightened activity in the insular cortex and regions within motor circuitry of adolescents indicates an affective-motivational sensitivity toward any peer approval. There also seem to be differences between different adolescent’s ages; the naturally occurring peak falls in late adolescence. In sum, there is thus an adolescent-specific effect of positive social feedback from peers on learning and neural activation patterns.