Collective Problem Solving in Guinea Baboons: From Dyadic Interactions to Group Dynamics (M/F)

Job Description

Thesis Subject

Project context and background
The fact that humans live in large groups is thought to have played a key role in the success of our species, as group life can bring important benefits (hunting of large preys, division of labour, cooperative care of infants, etc). To reap such benefits, individuals need to coordinate their actions (eg, when carrying loads), to cooperate (eg, in the defence against predators), or share valuable information such as tools and techniques. However, group living also comes with costs due to increased competition over food resources and mating opportunities. This tension creates an evolutionary pressure to develop mechanisms that reduce costs and increase benefits [1]. Humans use various strategies to develop, stabilize, and sometimes enforce cooperation: when confronted with a non-cooperative partner, humans may try to find a more cooperative one (partner choice strategies [2]). and/or try to increase the cooperativeness of the partner (partner control strategies), eg, using punishment [3] or a “Tit-for-Tat” strategy [4].
In the case of non-human primates, the strategies used and their role in the group dynamics are much less understood. One of the co-supervisors has performed preliminary studies on how non-human primates solve coordination and cooperation dilemmas [5,6], showing that Guinea baboons can adopt flexible strategies adapted to the problem they face. For instance, pairs of baboons had to choose between two stimuli randomly picked from a set. They could see each other and each other's response and were rewarded only when they chose the same stimuli (coordination task). In this context, individuals developed a response shared across the group. Moreover, results show that baboons that are central in the social network flexibly adapt to their neighbors' behaviour but not vice-versa. This suggests the existence of a complex dynamical process involving the social organization of the group.
What is still crucially lacking is an understanding of the link between the social abilities of non-human primates in dyadic interactions and their dynamics at the group level. For instance, both human and non-human primates cooperate to harvest hard to access resources. When pairs of individuals decide to cooperate, the outcome of the cooperation will influence the likelihood that the same individuals cooperate again, and the strategies used by two individuals will in turn affect other individuals in the group [7]. If two partners cooperate, this limits the opportunities for others to interact with them, creating complex group dynamics that depend on the relationships between individuals and impact their success at collectively solving the task.

Project goal and challenges
The objective of the PhD will be to provide a comprehensive analysis of the links between dyadic interactions and group dynamics for non-human primates during collective problem solving. The challenges are both experimental and methodological: Firstly, most non-human primate studies report only dyadic interactions because they do not have the means to study an entire group of freely interacting individuals. Secondly, modelling and understanding the mechanisms of the emergence of global effects in a social network from dyadic interactions is a challenging task, especially when feedback effects between dynamics and structure lead to a temporal evolution of the network.
The proposed co-supervision is uniquely positioned to enable the PhD student to tackle these challenges. N.C. has developed a unique computerised system in which a group of baboons can freely interact in pairs in various tasks. A.B. is an expert in the study of social networks and dynamical processes. A.B. and N.C. have co-supervised a PhD student in the past, leading to three research articles on the baboons' social network and its evolution.

Description of work and role of the PhD student
We expect non-human primates to show a variety of individual strategies to solve collective problems (e.g., the capacity to choose partners, to reciprocate, etc.), and individuals to flexibly adapt their strategies depending on the behaviour of other individuals in their group and the problem they are facing. To understand the effect of these strategies on group dynamics, we will use a social network approach in which dyadic interactions form the constituent elements of a social network, with each interaction impacting the social relationships of individuals [8] and the likelihood that these individuals will interact again. The evolution of dyadic interactions influences the rest of the network of social relationships when individuals change partners or adapt their behaviour to others. Here, the goal is also to understand how individuals adapt to the different situations and the benefits or costs that come from the group dynamics. We will moreover compare different types of collective problems (cooperation, coordination…) and investigate if and how global coordination emerge, comparing with various models that have been proposed to describe such emergence.

Experimental work. Experiments will take place at the Primate research station of Rousset-sur-Arc, where a mixed group of ~20 Guinea baboons (Papio papio) have freely access to computerized workstations. The voluntary participation of the subjects reduces stress and this platform allows a precise study of individual and social cognition [9]. Guinea baboons have a multilevel social system and the group that we will study is composed of 4-5 smaller social units. This social structure is adapted to our project because the baboons hierarchical network resembles human social networks. Moreover, we have experience studying and describing this network [10].
Baboons will repeatedly interact in pairs and will have to choose stimuli on a touchscreen to obtain rewards (wheat grains), with different conditions corresponding to (i) coordination (the two need to choose the same stimulus to get a reward) (ii) cooperation (one participant can choose between one reward for themselves only or one reward for both) (iii) competition (one has to choose the same stimuli as the other while the second individual is rewarded if different stimuli are chosen). To ascertain the social aspect of the strategies used by the participants, we will perform a “ghost” control condition in which the participants are not paired with a partner but with a computer agent that behaves as a partner. The contrast between the control and test conditions informs us on the strategies that are developed in a social context, compared to those simply linked to the task.

Analysis and modelling. We will use statistical physics tools to analyze the strategies used by individuals and how they create a complex social network dynamics. The data will consist in temporal networks of interactions between participants, together with individual choices and rewards. We will first analyse separately networks and rewards and then verify whether correlations exist between networks' features and measures of success (both individual and global). We will analyze the network both at the static (aggregated over time) and temporal level, leveraging the network science tools developed in the last decade, and compare the networks and task dynamics for the different conditions [11]. We will moreover consider various agent-based models, developed in statistical physics to describe the emergence of coordination/cooperation/competition in a population, usually studied on a non-evolving structure of interactions. We will extend them with individual strategies for changing interaction partners, and compare their dynamics with the observations. To compare the dynamics observed in different conditions and to compare them with models, we will develop new theoretical tools: indeed, while measures of differences between networks have been defined, and while one can compare the final outcome of different processes on networks, we lack a principled way to compare quantitatively, in detail and across several timescales, the observed dynamics with model dynamics in such a complex situation of feedback between structure and dynamics.

The role of the PhD student will be as follows:
-to contribute to the design of the experiments and to implement the experiments
-to analyse the collected data on the evolution of the networks and on the dynamics at the individual and global scale, to characterize the observed networks, investigate their evolution mechanisms and compare the networks in different conditions;
-to perform a literature review on models describing the emergence of coordination and cooperation, and the case of competition; to design extensions of the models taking into account individual strategies and to implement numerical simulations;
-to develop methods to compare the experimental dynamics in various conditions and with models;
-to propose and design new experiments inspired by the models' results.

Profile of the PhD candidate
The candidate should have broad interests spanning biological sciences, social sciences, mathematical modelling, numerical computing.
A strong expertise in statistical physics, modelling, numerical simulations, data analysis will be required (e.g., an engineering background from a grande école or an international master's in complex systems). Moreover, the candidate should be interested in working at the interface with social sciences and in designing and conducting experiments on animal behaviour.
Such a profile is essential to tackle the two complementary aspects of the PhD, namely on the one hand conducting experiments and behavioural analyses, and on the other hand developing and testing complex mathematical models that capture the dynamics of the networks and the one taking place on networks.


References: 1. Cooney, D.B., et al., Evolutionary dynamics within and among competing groups. PNAS, 2023. 120:e2216186120. 2. Barclay, P., Strategies for cooperation in biological markets, especially for humans. Evolution and Human Behavior, 2013. 34:164-175. 3. Fehr, E. and U. Fischbacher, The nature of human altruism. Nature, 2003. 425:785-791. 4. Rand, D.G. and M.A. Nowak, Human cooperation. Trends in Cognitive Sciences, 2013. 17:413-425. 5. Formaux, A., et al., The experimental emergence of convention in a non-human primate. Phil Trans Royal Soc B: Biological Sciences, 2021. 377: 20200310. 6. Formaux, A., et al., Guinea baboons are strategic cooperators. Sci Adv, 2023. 9(43): p. eadi5282. 7. Kozma, B. and Barrat, A., Consensus formation on coevolving networks: groups' formation and structure. J Phys A: Mathematical and Theoretical, 2008. 41:224020. 8. Gelardi, V., et al., From temporal network data to the dynamics of social relationships. Proc Royal Soc B: Biological Sciences, 2021. 288: 20211164. 9. Fagot, J., et al., Assessment of Social Cognition in Non-human Primates Using a Network of Computerized Automated Learning Device (ALDM) Test Systems. JoVE, 2015. 99:e52798. 10. Gelardi, V., et al., Measuring social networks in primates: wearable sensors versus direct observations. Proc Royal Soc A: Mathematical, Physical and Engineering Sciences, 2020. 476:20190737. 11. Dall'Amico, L., A. Barrat, and C. Cattuto, An embedding-based distance for temporal graphs. Nat. Commun 2024,.15, 9954.

Your Work Environment

The PhD will take place partly at the Research Center for Psychology and Neuroscience (3 Place Victor Hugo, 13331 Marseille) and at the Center for Theoretical Physics (163 Av. de Luminy, 13009 Marseille), where Nicolas Claidière and Alain Barrat are respectively affiliated. The experimental component will be conducted at the Primatology Research Station in Rousset (13790).

Compensation and benefits

Compensation

2300 € gross monthly

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions