Quantifying diffusion in social networks: a Bayesian approach

Prima facie, the acquisition of novel behaviors in animals through observation of conspecifics seems straightforward. There are, however, various mechanisms through which the behavior of animals can be altered from observing others. These mechanisms range from simple hard-wired contagious processes to genuine learning by observation, which differ fundamentally in cognitive complexity. They range from social facilitation and local enhancement to true social learning. The different learning mechanisms are the subject of this review, largely because research on learning by observation can be confounded by difficulties in interpretation owing to the looming possibility of associative learning infecting experimental results. While it is often assumed that horses are capable of acquiring new behavior through intra-species observation, research on social learning in horses includes a variety of studies some of which may overestimate the possession of higher mental abilities. Assuming such abilities in their absence can have welfare implications, e.g., isolating stereotypical horses on the assumption that these behaviors can be learned though observation by neighboring horses. This review summarizes the definitions and criteria for the various types of social transmission and social learning and reviews the current documentation for each type in horses with the aim of clarifying whether horses possess the ability to learn through true social learning. As social ungulates, horses evolved in open landscapes, exposed to predators and grazing most of the day. Being in close proximity to conspecifics may theoretically offer an opportunity to learn socially, however anti-predator vigilance and locating forage may not require the neural complexity of social learning. Given the significant energetic expense of brain tissue, it is likely that social facilitation and local enhancement may have been sufficient in the adaptation of equids to their niche. As a consequence, social learning abilities may be maladaptive in horses. Collectively, the review proposes a novel differentiation between social transmission (social facilitation, local, and stimulus enhancement) and social learning (goal emulation, imitation). Horses are undoubtedly sensitive to intra-species transfer of information but this transfer does not appear to satisfy the criteria for social learning, and thus there is no solid evidence for true social learning in horses.

There has been relatively little research on the psychological mechanisms of social learning. This may be due, in part, to the practice of distinguishing categories of social learning in relation to ill-defined mechanisms (Davis, 1973; Galef, 1988). This practice both makes it difficult to identify empirically examples of different types of social learning, and gives the false impression that the mechanisms responsible for social learning are clearly understood. It has been proposed that social learning phenomena be subsumed within the categorization scheme currently used by investigators of asocial learning. This scheme distinguishes categories of learning according to observable conditions, namely, the type of experience that gives rise to a change in an animal (single stimulus vs. stimulus-stimulus relationship vs. response-reinforcer relationship), and the type of behaviour in which this change is detected (response evocation vs. learnability) (Rescorla, 1988). Specifically, three alignments have been proposed: (i) stimulus enhancement with single stimulus learning, (ii) observational conditioning with stimulus-stimulus learning, or Pavlovian conditioning, and (iii) observational learning with response-reinforcer learning, or instrumental conditioning. If, as the proposed alignments suggest, the conditions of social and asocial learning are the same, there is some reason to believe that the mechanisms underlying the two sets of phenomena are also the same. This is so if one makes the relatively uncontroversial assumption that phenomena which occur under similar conditions tend to be controlled by similar mechanisms. However, the proposed alignments are intended to be a set of hypotheses, rather than conclusions, about the mechanisms of social learning; as a basis for further research in which animal learning theory is applied to social learning. A concerted attempt to apply animal learning theory to social learning, to find out whether the same mechanisms are responsible for social and asocial learning, could lead both to refinements of the general theory, and to a better understanding of the mechanisms of social learning. There are precedents for these positive developments in research applying animal learning theory to food aversion learning (e.g. Domjan, 1983; Rozin & Schull, 1988) and imprinting (e.g. Bolhuis, de Vox & Kruit, 1990; Hollis, ten Cate & Bateson, 1991). Like social learning, these phenomena almost certainly play distinctive roles in the antogeny of adaptive behaviour, and they are customarily regarded as 'special kinds' of learning (Shettleworth, 1993).(ABSTRACT TRUNCATED AT 400 WORDS)

Cultured cetaceans

Social learning has been documented in a wide diversity of animals. In free-living animals, however, it has been difficult to discern whether animals learn socially by observing other group members or asocially by acquiring a new behaviour independently. We addressed this challenge by developing network-based diffusion analysis (NBDA), which analyses the spread of traits through animal groups and takes into account that social network structure directs social learning opportunities. NBDA fits agent-based models of social and asocial learning to the observed data using maximum-likelihood estimation. The underlying learning mechanism can then be identified using model selection based on the Akaike information criterion. We tested our method with artificially created learning data that are based on a real-world co-feeding network of macaques. NBDA is better able to discriminate between social and asocial learning in comparison with diffusion curve analysis, the main method that was previously applied in this context. NBDA thus offers a new, more reliable statistical test of learning mechanisms. In addition, it can be used to address a wide range of questions related to social learning, such as identifying behavioural strategies used by animals when deciding whom to copy.

Culture is an outcome of both the acquisition of knowledge about behaviour through social transmission, and its subsequent production by individuals. Acquisition and production are often discussed or modelled interchangeably, yet to date no study has explored the consequences of their interaction for cultural diffusions. We present a generative model that integrates the two, and ask how variation in production rules might influence diffusion dynamics. Agents make behavioural choices that change as they learn from their productions. Their repertoires may also change, and the acquisition of behaviour is conditioned on its frequency. We analyse the diffusion of a novel behaviour through social networks, yielding generalizable predictions of how individual-level behavioural production rules influence population-level diffusion dynamics. We then investigate how linking acquisition and production might affect the performance of two commonly used inferential models for social learning; network-based diffusion analysis, and experience-weighted attraction models. We find that the influence that production rules have on diffusion dynamics has consequences for how inferential methods are applied to empirical data. Our model illuminates the differences between social learning and social influence, demonstrates the overlooked role of reinforcement learning in cultural diffusions, and allows for clearer discussions about social learning strategies.

Microbial transmission in the social microbiome and host health and disease

The emergence and spread of novel behaviours via social learning can lead to rapid population-level changes whereby the social connections between individuals shape information flow. However, behaviours can spread via different mechanisms and little is known about how information flow depends on the underlying learning rule individuals employ. Here, comparing four different learning mechanisms, we simulated behavioural spread on replicate empirical social networks of wild great tits and explored the relationship between individual sociality and the order of behavioural acquisition. Our results reveal that, for learning rules dependent on the sum and strength of social connections to informed individuals, social connectivity was related to the order of acquisition, with individuals with increased social connectivity and reduced social clustering adopting new behaviours faster. However, when behavioural adoption depends on the ratio of an individuals’ social connections to informed versus uninformed individuals, social connectivity was not related to the order of acquisition. Finally, we show how specific learning mechanisms may limit behavioural spread within networks. These findings have important implications for understanding whether and how behaviours are likely to spread across social systems, the relationship between individuals’ sociality and behavioural acquisition, and therefore for the costs and benefits of sociality.

The emergence and spread of novel behaviours via social learning can lead to rapid population-level changes whereby the social connections between individuals shape information flow. However, behaviours can spread via different mechanisms and little is known about how information flow depends on the underlying learning rule individuals employ. Here, comparing four different learning mechanisms, we simulated behavioural spread on replicate empirical social networks of wild great tits and explored the relationship between individual sociality and the order of behavioural acquisition. Our results reveal that, for learning rules dependent on the sum and strength of social connections to informed individuals, social connectivity was related to the order of acquisition, with individuals with increased social connectivity and reduced social clustering adopting new behaviours faster. However, when behavioural adoption depends on the ratio of an individuals' social connections to informed versus uninformed individuals, social connectivity was not related to the order of acquisition. Finally, we show how specific learning mechanisms may limit behavioural spread within networks. These findings have important implications for understanding whether and how behaviours are likely to spread across social systems, the relationship between individuals' sociality and behavioural acquisition, and therefore for the costs and benefits of sociality.

A critical consequence of joining social groups is the possibility of social transmission of information related to novel behaviours or resources. Network‐based diffusion analysis (NBDA) has emerged as a leading frequentist framework for inferring and quantifying social transmission, particularly in non‐human animal populations. NBDA has been extended several times to account for multiple diffusions, multiple networks, individual‐level variables and complex transmission functions. Bayesian versions of NBDA have been proposed before, although these implementations have seen limited usage and have not kept pace with the evolving ecosystem of Bayesian methods. There is not yet a user‐friendly package to implement a Bayesian NBDA. Here, we present a unified framework for performing Bayesian analysis of social transmission using NBDA‐type models, implemented in the widely used Stan programming language. We provide a user‐friendly R package ‘STbayes’ (ST: social transmission) for other researchers to easily use this framework. STbayes accepts user‐formatted data, but can also import data directly from the existing NBDA R package. Based on the data users provide, STbayes automatically generates multi‐network, multi‐diffusion models that allow for covariates that may influence transmission and varying (random) effects. Using simulated data, we demonstrate that this model can accurately differentiate the relative contribution of individual and social learning in the spread of information through networked populations. We illustrate how incorporating upstream uncertainty about the relationships between individuals can improve model fit. Our framework can be used to infer complex transmission rules, and we describe a numerically stable parametrization of frequency‐dependent transmission. Finally, we introduce support for dynamic transmission weights and a ‘high‐resolution’ data mode, which allows users to make use of fine‐scale data collected by contemporary automated tracking methods. These extensions increase the set of contexts that this type of model may be used for.

Familial and social transmission of suicidal behavior in older adults

Spread of two linked social norms on complex interaction networks

Although social learning capabilities are taxonomically widespread, demonstrating that freely interacting animals (whether wild or captive) rely on social learning has proved remarkably challenging. Network-based diffusion analysis (NBDA) offers a means for detecting social learning using observational data on freely interacting groups. Its core assumption is that if a target behaviour is socially transmitted, then its spread should follow the connections in a social network that reflects social learning opportunities. Here, we provide a comprehensive guide for using NBDA. We first introduce its underlying mathematical framework and present the types of questions that NBDA can address. We then guide researchers through the process of selecting an appropriate social network for their research question; determining which NBDA variant should be used; and incorporating other variables that may impact asocial and social learning. Finally, we discuss how to interpret an NBDA model's output and provide practical recommendations for model selection. Throughout, we highlight extensions to the basic NBDA framework, including incorporation of dynamic networks to capture changes in social relationships during a diffusion and using a multi-network NBDA to estimate information flow across multiple types of social relationship. Alongside this information, we provide worked examples and tutorials demonstrating how to perform analyses using the newly developed nbda package written in the R programming language.

Previous empirical work on animal social learning has found that many species lack the ability to learn entire action sequences solely through reliance on social information. Conversely, acquiring action sequences through asocial learning can be difficult due to the large number of potential sequences arising from even a small number of base actions. In spite of this, several studies report that some primates use action sequences in the wild. We investigate how social information can be integrated with asocial learning to facilitate the learning of action sequences. We formalize this problem by examining how learners using temporal difference learning, a widely applicable model of reinforcement learning, can combine social cues with their own experiences to acquire action sequences. The learning problem is modeled as a Markov decision process. The learning of nettle processing by mountain gorillas serves as a focal example. Through simulations, we find that the social facilitation of component actions can combine with individual learning to facilitate the acquisition of action sequences. Our analysis illustrates that how even simple forms of social learning, combined with asocial learning, generate substantially faster learning of action sequences compared to asocial processes alone, and that the benefits of social information increase with the length of the action sequence and the number of base actions.

Prosocial effects of prolactin in male rats: Social recognition, social approach and social learning

Human cognition is distinguished by our ability to adapt to different environments and circumstances. Yet the mechanisms driving adaptive behavior have predominantly been studied in separate asocial and social contexts, with an integrated framework remaining elusive. Here, we use a collective foraging task in a virtual Minecraft environment to integrate these two fields, by leveraging automated transcriptions of visual field data combined with high-resolution spatial trajectories. Our behavioral analyses capture both the structure and temporal dynamics of social interactions, which are then directly tested using computational models sequentially predicting each foraging decision. These results reveal that adaptation mechanisms of both asocial foraging and selective social learning are driven by individual foraging success (rather than social factors). Furthermore, it is the degree of adaptivity—of both asocial and social learning—that best predicts individual performance. These findings not only integrate theories across asocial and social domains, but also provide key insights into the adaptability of human decision-making in complex and dynamic social landscapes.