Competitive and cooperative games for probing the neural basis of social decision-making in animals

Abstract

In a social environment, it is essential for animals to consider the behavior of others when making decisions. To quantitatively assess such social decisions, games offer unique advantages. Games may have competitive and cooperative components, modeling situations with antagonistic and shared objectives between players. Games can be analyzed by mathematical frameworks, including game theory and reinforcement learning, such that an animal's choice behavior can be compared against the optimal strategy. However, so far games have been underappreciated in neuroscience research, particularly for rodent studies. In this review, we survey the varieties of competitive and cooperative games that have been tested, contrasting strategies employed by non-human primates and birds with rodents. We provide examples of how games can be used to uncover neural mechanisms and explore species-specific behavioral differences. We assess critically the limitations of current paradigms and propose improvements. Together, the synthesis of current literature highlights the advantages of using games to probe the neural basis of social decisions for neuroscience studies.

Keywords: Behavioral strategy; Game theory; Matching pennies; Reinforcement learning; Social behavior; Social neuroscience.

Figure 1.. Social games used in animal decision-making.

(A) Payoff matrix for matching pennies. a,b,c,d denote the amount of outcome gained or lost depending on the actions. (B) Payoff matrix for rock-paper-scissors, a denotes the amount of outcome. Blue and red text indicating the actions and outcomes for the corresponding players. (C) Payoff matrix for inspection game. i denotes the cost of inspection. c > g. (D) Payoff matrix for prisoner’s dilemma. t > r > p > s. (E) Payoff matrix for assurance game. a > b > d > c. (F) Payoff matrix for snowdrift game. t > 1 > s > 0. (G) Extensive form representation for ultimatum game. The proposer offers the responder y, while the total outcome is x. If the responder accepts, the proposer gets x-y, the responder gets y. If the responder rejects, both players get 0. (H) Extensive form representation for the dictator game. The proposer offers the responder y out of x, which is the final decision. (I) Extensive form representation for trust game. The trustor gives a proportion (p) of the reward X, which is multiplied by K when the trustee gets it. Then the trustee gives another proportion (q) back to the trustor. Both p and q can be 0.

Figure 2.. Performance of monkeys and mice in matching pennies.

(A) Behavior paradigm for monkey and mouse. Top: monkeys use saccade to indicate the decisions on computer screen. Bottom: mice use tongue-licking to indicate the decisions and get water reward from the licking port. (B) Psychometric curve of the behavior of monkeys (Lee et al., 2004) and mice (Wang et al., 2022) determined by fitting similar reinforcement learning-based computational models. Orange histograms represent the distribution of trials according to the difference of action values, which are variables related to action selection. Black curves represent the predicted behavior by the model. Black dots represent the observed behavior in actual experiment. (C) Model estimations of learning rate (latent parameter that modulates the strength of influence of most recent outcome on behavior) and inverse temperature (modulates the variability of behavior) of monkey data (Seo et al., 2009) and mouse data (Wang et al., 2022). (D) The probability of animals adopted win-stay-loose-switch strategy plotted against the estimated learning rate of monkey data (Seo et al., 2009) and mouse data (Wang et al., 2022). (E) Logistic regression coefficients relating the reinforcer effect bias with the current choice of monkey data (Seo et al., 2009) and mouse data (Wang et al., 2022). (F) Logistic regression coefficients relating the choice history bias with the current choice of monkey data (Seo et al., 2009) and mouse data (Wang et al., 2022).

Figure 3.. Brain regions related to social games and strategic behavior.

(A) Lateral view of macaque brain with regions involved in competitive game (yellow) and cooperative game (blue). ACC: anterior cingulate cortex. DLPFC/DMPFC: dorsolateral/dorsomedial prefrontal cortex. LIP: lateral intraparietal cortex. mSTS: medial superior temporal sulcus. SC: superior colliculus. SEF: supplement eye field. (B) Sagittal view of mouse brain with regions involved in competitive game. ACC: anterior cingulate cortex. BF: basal forebrain. LC: locus coeruleus. STR: striatum.