What's the science?
Leadership is critical within our society. Leaders such as teachers, soldiers, politicians and parents, to name a few, are continuously responsible for making decisions that will affect others. One aspect of leadership is the acceptance of responsibility, as leaders are responsible for others in the choices they make. Despite the fact that leadership is central to human societies, we still don’t understand the neurobiology of leadership and why some choose to lead and others chose to follow. This week in Science, Edelson and colleagues use a decision-making task to examine leadership choices and the brain regions involved.
How did they do it?
They developed a behavioural task to examine leadership preferences. The participants performed 2 tasks: a baseline task and a delegation task. In all tasks, the participant’s payoff was dependent on their choices, and therefore assessed preferences related to risk, loss and ambiguity. In the baseline task, they were required to choose whether they would accept a gamble with varying probabilities of loss or gain over several trials. In some trials the exact probability of gains and losses were shown to the participant (to assess risk), while in some trials these probabilities were not shown (i.e. ambiguous and closer to real life decisions). In the delegation task, participants were required to decide on the same gambles, however, they could choose to lead and make a decision on behalf of their group, or defer and follow the choice of the group members. In this task there were two trials, ‘self’ and ‘group’ trials, where the choice of the leader would directly affect the payoff of him or herself or affect the payoff of the group members. The group members as a whole had more information about the probabilities of an outcome, which mimics a real-life scenario where deciding as a group may be advantageous. The authors analyzed baseline choice data to determine whether preferences for risk, loss or ambiguity were associated with leadership scores (obtained using established scales, as well as real life data) and whether there is a shift in these preferences when making choices that impact others. They then used computational modelling and fMRI analysis to understand the preferences for leadership and the brain activity underlying these choices.
What did they find?
In the ‘group’ trials, participants deferred to the decision of the group more often compared to the ‘self’ trials (17.3% increase in deferral rate), demonstrating an overall preference for avoiding responsibility. Responsibility aversion (or avoidance of responsibility) was correlated with leadership scores, where those with lower responsibility aversion had higher leadership scores. Responsibility aversion was not related to individual preferences/ values related to risk, loss and ambiguity, nor was it related to regret, blame, and guilt in social situations (assessed in a separate experiment). They then used a computational model taking into account the participants preferences to try and model these decisions. From participants preferences they were able to derive subjective values (i.e. difference in value between accepting or rejecting a gamble for each participant) of choices to lead or defer. They found that deferral choices were greater when there was a low subjective value difference (i.e. less discriminability between the value of the two choices). In other words, when the participant was more certain of their choice (high discriminability), they were more likely to take on responsibility. They derived ‘deferral thresholds’ whereby a critical level of certainty decides whether they will defer or not in a gamble. Using computational modelling they found that when taking responsibility for others there was a shift in the deferral threshold, indicating a higher demand for certainty of a choice. These changes in deferral threshold were correlated with leadership scores. The greater the shift in the deferral threshold, the more likely an individual was to defer (i.e. not lead). Since individual values about risk, loss and ambiguity were not related to leadership scores, these results suggest that the key to determining whether one will lead or not lies in the shift in the amount of certainty needed (about their decision) now that they are responsible for others.
Brain activation was higher in the middle temporal gyrus during group trials compared to self-trials, and temporo-parietal junction activity was correlated with the decision to defer (i.e. not lead). Activity in several brain regions including the medial prefrontal cortex (involved in self-reflection) was associated with the subjective value difference while activity in the anterior insula was associated with the probability of choosing to lead. The authors fit a dynamic causal model to these four brain regions and found a relationship between individual differences in activity in this brain network, shift in deferral threshold and leadership scores. Temporal gyrus activity was also associated with a reduced or inhibited influence of the medial prefrontal cortex on anterior insula activity. This inhibitory influence of medial prefrontal cortex on the anterior insula was reduced in leaders, suggesting a potential neural mechanism underlying the shift in deferral threshold and responsibility aversion.
What's the impact?
This is the first study to examine the behaviour and neurobiology underlying leadership preferences. We now know that the majority of individuals show responsibility aversion and that this is a critical driver in the decision to lead. Those who were less likely to lead show a greater difference in their demand for certainty when deciding to take responsibility for others. Further, this study provides important insight into the brain regions involved in leadership choices. These findings have important implications for understanding leadership in society and can help to inform leadership decisions and consequences.
Edelson et al., Computational and neurobiological foundations of leadership decisions. Science (2018). Access the original scientific publication here.