It’s no secret that social factors significantly influence our choices and our learning.  Indeed, the notion of social influence was first coined in the 1950s by Solomon Asch.  New research from the University Medical Center Hamburg-Eppendorf (UKE) aims to add its own perspective to the topic.

It saw groups of volunteers placed into a computer-based decision-making experiment whereby they were presented with a couple of abstract symbols.  The aim was to understand which symbol would be linked with the highest monetary rewards over the long-run.  In each round of the experiment the volunteer would make a choice between the two symbols, before then also observing the choices the other four people in their group made.

They were then given the choice of whether to stick with their choice or change.  After everyone had decided on their choice, they were told whether they won or lost money as a result.  The trick was that the symbol linked with the high reward was changing constantly.

“This so-called reversal learning paradigm will create uncertainty for volunteers so that they will always need to learn and relearn to gain more outcomes. In particular, when the reversal just happened, some people in the group may pick it faster than the others, and if so, the others could combine this social information into their own decision-making processes,” the researchers explain.

Uncertain choices

As expected, people tended to switch their choices more often when they were exposed to divergent choices in their group.  What was interesting, however, is that this second choice tended to better reflect the reward structure in the game.

The researchers attempted to explain this via sophisticated models that allowed them to quantify the behavior of the volunteers.

“At the beginning of each round, the volunteers were combining their own direct learning experience and social learning experience to guide their choice,” the researchers explain, “whereby direct learning follows a simple reinforcement learning algorithm, and social learning is instantiated by tracking the others’ reward history.”

The brains of the volunteers were scanned using functional magnetic resonance imaging, which allowed the researchers to measure when and where the brain was carrying out both direct and social learning.  Indeed, were the two forms of learning associated with different ‘neural signatures’?

These scans revealed that directional learning appeared to take place in the ventromedial prefrontal cortex, with the social learning undertaken in the anterior cingulate cortex.  Both of these areas would also interact with the striatum, which is used to predict reward error, thus allowing us to quantify trial and error learning.

So the two forms of learning are computed in distinct, yet interacting, parts of the brain, with distinct computational strategies at play when making decisions in social contexts.

“Direct learning is efficient in stable situations,” the authors conclude, “and when situations are changing and uncertain, social learning may play an important role together with direct learning to adapt to novel situations, such as deciding on the lunch menu at a new company.”