What is the "Collective Prisoner's Dilemma"?

Brain science reveals how groups make shared decisions.

Posted Aug 16, 2020

How we make decisions together when we are looking to maximize the benefit to all is an important area of research, and important for the survival of our species. We’re passing the point where we can act as if we are competing on a playing field with infinite resources. 

Planetary prisoners

Nations are approaching full interdependence as they expand to take up as much space as they can on the global landscape. Space is another frontier, but on the surface of the planet, we’re shoulder-to-shoulder.

Our situation is akin to what game theorists call the “repeated Prisoner’s Dilemma” (PD). In the single-case PD, co-conspirators are captured, then separated for interrogation. If they keep the faith (cooperate), they each get a moderate reward, and go free. If they stab each other in the back (defect), they both get punished. If one defects, and the other cooperates, the defector gets a sweet deal, and the other... well, you know. Go directly to jail, do not pass Go, do not collect $200.

 Christopher X Jon Jensen (CXJJensen) & Greg Riestenberg, Wikimedia Commons Open Source
The Prisoner's Dilemma Illustrated
Source: Christopher X Jon Jensen (CXJJensen) and Greg Riestenberg, Wikimedia Commons Open Source

This paradigm is used as a way of studying how we make decisions, part of "game theory." Do nations cooperate, or defect? Are short-term strategies, which may favor defection, good in the long-run?

The long view

Research on the best long-term strategies for repeated PD support ongoing cooperation as the winning approach. Cooperation on balance works out best for all, even with some defections. Illustrating this, in a repeated PD study of 94 participants playing 400 10-round games, researchers found that 40 percent of players were "resilient cooperators." Based on this data, modelling predicted that “a sufficiently large minority of resilient cooperators can permanently stabilize unravelling among a majority of rational players.” That's how to win the "collective prisoner's dilemma". 

Individual factors

In their paper "Group decision-making behavior in social dilemmas: Inter-brain synchrony and the predictive role of personality traits," Zhang, Jia, Zheng and Liu (2020) set-up a variant of the PD, pitting two players against one (who was actually one of the researchers) in order to study how Big 5 personality factors (Openness to Experience, Conscientiousness, Extraversion, Agreeableness, Neuroticism) and brain activity correlate with decision-making.

They recruited 54 participants who did not previously know one another into 27 pairs. Before the PD games, they completed a measure of Big 5 personality traits. While they were playing, participants’ brain activity was captured with fNIRS (functional near-infrared spectroscopy, with a focus on key areas identified from prior research. It's easier than fMRI for studies like this as players don't need to be in a scanner.

There were two parts of each game. In the first, participants decided to cooperate or defect without conferring (the individual decision-making stage, IDM). In the second, they discussed their strategy (the group decision-making stage, GDM). This way they compared self-interest with shared interest.

Researchers also set high and low reward scenarios (high incentive reward, HIR, and low incentive reward, LIR) in order to see how the stakes affected findings. Brain activity was measured throughout, along with reaction times and outcomes.


Overall, there was more cooperation than defection with lower stakes. Players were faster at deciding what to do when cooperating; deciding to defect slowed them down. Cooperation was more common in group decision-making compared with individual.

The brain's right inferior frontal gyrus (rIFG), thought to be a key component in the human mirror neuron system (involved with empathy and attunement), was more active during GDM. Inter-brain synchrony (IBS) was higher during GDM, seen in increased rIFD coherence. Players' brains were more entrained together at those times. Their IFDs were also more synced-up when the reward was higher. Sync (IBS) spiked in the high-reward/group decision-making condition, suggesting that the mirror neuron system is most engaged when our fates are intertwined and the stakes are high.

IBS was elevated in the right dorsolateral prefrontal cortex (rDLPFC) during GDM. The rDLPFC is key in executive function, mediating cognitive control and moral decision-making. Greater effort may be required to make the best decision for all in exercising joint top-down control when the immediate impulse may be to go for the quick win.

Personality tracked with these findings. Extroversion and agreeableness correlated with IBS. Extroversion tracked with rIFG IBS during cooperative decisions, and agreeableness with lIFG. Agreeableness and extroversion correlated with rIFG IBS when deciding to cooperate in the HIR task. On the other hand, IBS in the DLPFC was lower when pairs chose defection, reflective of "being of two minds" about what to do.

For consideration

This work is preliminary, extending earlier work on group decision-making showing that we tend to be more rational when we think together about what is best for all. Personality plays a role—agreeableness and extroversion are connected with greater brain sync between players during cooperation with group decisions. Prior work has shown that agreeableness, extroversion, and conscientiousness are correlated with cooperative choices, on average.

The role of two important brain areas, the IFG and the DPLFC, were highlighted by this study. Given the IFG’s role in the mirror neuron system, it makes sense that group intelligence is reflected in sync (or lack thereof) between these areas. When people are working together collectively, our brains must be entrained, figuratively constituting the “hive mind” (which is fascinating from a neuroscience point of view1).

Are we selfish by default, or cooperative by default? Personality and upbringing make a difference, as does the need for rational group decision-making. People higher in “dark traits” are less likely to cooperate unless they are higher in empathy (“dark empaths”) or if there is clear self-interest and they are not so narcissistic, sadistic, or psychopathic they can’t act in their own self-interest when it also serves others’ needs. For people average or low on dark traits and average or higher on empathy, cooperation may be more or less an obviously good idea, depending on various factors including personality and upbringing. 

In terms of working out how to share the planet most effectively, making the cooperative long-term decisions which game theory suggests are the best strategy for future generations,2 it’s important to combined education with rational decision-making, using tools to strengthen the factors which enhance cooperation on for individuals and groups, and setting up reward systems which increased the odds of choosing the best overall long-term strategies.

Amplifying the voices of—and providing support and resources to—“resilient cooperators” who can shift toward positive outcomes even as a significant minority may be a winning strategy. Understanding who these people are and determining if they are best suited for leadership positions could be part of a rational strategy to secure a better future for humanity.


1. The DLPFC is interesting because it mediates top-down cognitive control especially when moral values are implicated in group decisions. In defection, there is a lack of sync between players’ DLPFCs. The time required to decide to defect is longer than the time required to cooperate, as we morally grapple with the decision to betray rather than the more prosocial move to cooperate.

The DLPFC, and closely related areas, are involved in many key functions. For instance, prior work suggests that we take in collective knowledge via the dorsomedial prefrontal cortex (DMPFC). Our tendency to pay attention to our own needs, a factor involved in weighing group decisions which may not immediately gratify, is mediated by the DMPFC as well, a key area in the brain’s activity at rest, or default mode network. We are programmed by our culture, and tend to focus on our own needs for basic survival--yet in order to survive in the long-run, we have to function collectively.

When it comes to social relations, the ventromedial prefrontal cortex (VMPFC) comes on line. Whether or not you’ll be friends three months after meeting someone can be predicted based on VMPFC activity at the time, both individual and in sync. Other work has pointed to the key role of the DLPFC in radicals, such as people who hold terrorist beliefs, and connects with the willingness to act on—even die for—those beliefs.

2. Science-fiction out-take: Perhaps decision-makers can be hooked-up to fNIRS and related systems, during negotiations. Real-time brain-activity can be used to gauge the degree of true synchrony during group decision-making tasks, and stake-holders (such as voters, the media, etc.) can pay attention to meaningful neuroscience activity. I, for one, would like to see streaming graphics on candidates' brain activity in order to inform my decision-making, at least.

Please note: An ExperiMentations Blog Post ("Our Blog Post") is not intended to be a substitute for professional advice. We will not be liable for any loss or damage caused by your reliance on information obtained through Our Blog Post. Please seek the advice of professionals, as appropriate, regarding the evaluation of any specific information, opinion, advice, or other content. We are not responsible and will not be held liable for third party comments on Our Blog Post. Any user comment on Our Blog Post that in our sole discretion restricts or inhibits any other user from using or enjoying Our Blog Post is prohibited and may be reported to Sussex Publishers/Psychology Today. Grant H. Brenner. All rights reserved.