Distinct Roles for Social Brain Network Regions in Strategy Development

Post by Lincoln Tracy

What's the science?

Social interactions lead to bursts of brain activity in the “social brain network”; a collection of different brain regions involved in social functioning. The right temporoparietal junction (rTPJ) is thought to play a crucial role in social-related brain activity. However, little is known about what the rTPJ and other brain regions do during these active periods, or if and how this activity differs depending on the specific social context. This week in Neuron, Konovalov and colleagues used changes in blood-oxygen-level-dependent (BOLD) activity of the “social brain network” during a standardized game paradigm to break down network activity into different contexts (e.g., social versus non-social, etc.).

How did they do it?

The authors recruited 60 volunteers aged 18 to 25 and randomly assigned these individuals to the social or non-social context to complete the standardized game paradigm inside a functional magnetic resonance imaging scanner. Participants in the social context were told they were playing a game of hide and seek against human opponents. In the social condition, the goal was to find a coin that could be hidden behind either a rock or a tree. Participants in the non-social context were told they were playing a guessing game where they needed to predict the next card drawn from a deck. Participants completed more than 200 guessing trials and scored or lost points depending on whether their guess was similar or different to their opponent, depending on the context. All participants actually played against two different computer algorithms—the learner and the sequencer. The learner algorithm kept track of the player’s play history, estimated the frequency of the player’s choices, and played the less frequent option. In contrast, the sequencer algorithm ignored the player’s choices and played a sequence that switched every two trials (e.g., tree-tree-rock-rock-tree-tree…).

The authors combined the behavioral choice data from the games with BOLD activity for the different brain regions to answer a series of questions. First, they tested whether the choices participants made against the two different algorithms led to different success rates between the social and non-social contexts. Second, they explored the strategies players used in the social context and whether these were the optimal strategies to beat the algorithms. Third, they asked whether “social brain network” activity differed between different contexts and algorithms encountered during the game. Finally, they explored unique functional roles of the “social brain network” regions. Specifically, they were interested in how the rTPJ interacted with other regions during the game.   

What did they find?

First, the authors found that participants performed better against the learner algorithm in the social context (compared to the non-social context) but that performance against the sequencer algorithm was the same regardless of context. These results suggest that the social context invoked a specific strategy that benefits when coming up against a reactive opponent. Second, they found that participant choices matched the optimal strategy 72% of the time. Third, they found that all brain regions were strongly tied to the outcome of the game, with more activation following a win than a loss. However, there was greater activation throughout the “social brain network” regions when the participant played against the learner algorithm compared to the sequencer algorithm. This pattern was highly similar to the analysis of the behavioral choice data. Finally, they found that connectivity between the rTPJ and other “social brain network” regions was increased when the participant won the game, confirming their hypothesis that the rTPJ communicates behaviorally relevant outcome information to connected brain regions.

What's the impact?

These results provide a new way of considering similar standard laboratory tasks that measure activity in “social brain network” regions where participants have to consider the mental states of other people. These findings provide crucial novel findings, as they support clear functional differences between the rTPJ and other social-related regions. The computational approach employed as part of this study could be used in clinical populations—such as autism spectrum disorder—to better understand neurocognitive characteristics within these populations.

Konovalov et al. Dissecting functional contributions of the social brain to strategic behavior. Neuron (2021). Access the original scientific publication here.