A tunable algorithm for collective decision-making

Federated inference and belief sharing

This paper concerns the distributed intelligence or federated inference that emerges under belief-sharing among agents who share a common world-and world model. Imagine, for example, several animals keeping a lookout for predators. Their collective surveillance rests upon being able to communicate their beliefs-about what they see-among themselves. But, how is this possible? Here, we show how all the necessary components arise from minimising free energy. We use numerical studies to simulate the generation, acquisition and emergence of language in synthetic agents. Specifically, we consider inference, learning and selection as minimising the variational free energy of posterior (i.e., Bayesian) beliefs about the states, parameters and structure of generative models, respectively. The common theme-that attends these optimisation processes-is the selection of actions that minimise expected free energy, leading to active inference, learning and model selection (a.k.a., structure learning). We first illustrate the role of communication in resolving uncertainty about the latent states of a partially observed world, on which agents have complementary perspectives. We then consider the acquisition of the requisite language-entailed by a likelihood mapping from an agent's beliefs to their overt expression (e.g., speech)-showing that language can be transmitted across generations by active learning. Finally, we show that language is an emergent property of free energy minimisation, when agents operate within the same econiche. We conclude with a discussion of various perspectives on these phenomena; ranging from cultural niche construction, through federated learning, to the emergence of complexity in ensembles of self-organising systems.

Alternative model systems for cognitive variation: eusocial-insect colonies

Understanding the origins and maintenance of cognitive variation in animal populations is central to the study of the evolution of cognition. However, the brain is itself a complex, hierarchical network of heterogeneous components, from diverse cell types to diverse neuropils, each of which may be of limited use to study in isolation or prohibitively challenging to manipulate in situ. Consequently, highly tractable alternative model systems may be valuable tools. Eusocial-insect colonies display emergent cognitive-like properties from relatively simple social interactions between diverse subunits that can be observed and manipulated while operating collectively. Here, we review the individual-scale mechanisms that cause group-level variation in how colonies solve problems analogous to cognitive challenges faced by brains, like decision-making, attention, and search.

The emergence of a collective sensory response threshold in ant colonies

SignificanceIn this study, we ask how ant colonies integrate information about the external environment with internal state parameters to produce adaptive, system-level responses. First, we show that colonies collectively evacuate the nest when the ground temperature becomes too warm. The threshold temperature for this response is a function of colony size, with larger colonies evacuating the nest at higher temperatures. The underlying dynamics can thus be interpreted as a decision-making process that takes both temperature (external environment) and colony size (internal state) into account. Using mathematical modeling, we show that these dynamics can emerge from a balance between local excitatory and global inhibitory forces acting between the ants. Our findings in ants parallel other complex biological systems like neural circuits.

The Power of Population Effect in Temnothorax Ant House-Hunting: A Computational Modeling Approach

The decentralized cognition of animal groups is both a challenging biological problem and a potential basis for bioinspired design. In this study, we investigated the house-hunting algorithm used by emigrating colonies of Temnothorax ants to reach consensus on a new nest. We developed a tractable model that encodes accurate individual behavior rules, and estimated our parameter values by matching simulated behaviors with observed ones on both the individual and group levels. We then used our model to explore a potential, but yet untested, component of the ants' decision algorithm. Specifically, we examined the hypothesis that incorporating site population (the number of adult ants at each potential nest site) into individual perceptions of nest quality can improve emigration performance. Our results showed that attending to site population accelerates emigration and reduces the incidence of split decisions. This result suggests the value of testing empirically whether nest site scouts use site population in this way, in addition to the well-demonstrated quorum rule. We also used our model to make other predictions with varying degrees of empirical support, including the high cognitive capacity of colonies and their rational time investment during decision-making. In addition, we provide a versatile and easy-to-use Python simulator that can be used to explore other hypotheses or make testable predictions. It is our hope that the insights and the modeling tools can inspire further research from both the biology and computer science community.

Ants resort to majority concession to reach democratic consensus in the presence of a persistent minority

Social groups often need to overcome differences in individual interests and knowledge to reach consensus decisions. Here, we combine experiments and modeling to study conflict resolution in emigrating ant colonies during binary nest selection. We find that cohesive emigration, without fragmentation, is achieved only by intermediate-sized colonies. We then impose a conflict regarding the desired emigration target between colony subgroups. This is achieved using an automated selective gate system that manipulates the information accessible to each ant. Under this conflict, we find that individuals concede their potential benefit to promote social consensus. In particular, colonies resolve the conflict imposed by a persistent minority through "majority concession," wherein a majority of ants that hold first-hand knowledge regarding the superior quality nest choose to reside in the inferior one. This outcome is unlikely in social groups of selfish individuals and emphasizes the importance of group cohesion in eusocial societies.

TNF controls a speed-accuracy tradeoff in the cell death decision to restrict viral spread

Rapid death of infected cells is an important antiviral strategy. However, fast decisions that are based on limited evidence can be erroneous and cause unnecessary cell death and subsequent tissue damage. How cells optimize their death decision making strategy to maximize both speed and accuracy is unclear. Here, we show that exposure to TNF, which is secreted by macrophages during viral infection, causes cells to change their decision strategy from "slow and accurate" to "fast and error-prone". Mathematical modeling combined with experiments in cell culture and whole organ culture show that the regulation of the cell death decision strategy is critical to prevent HSV-1 spread. These findings demonstrate that immune regulation of cellular cognitive processes dynamically changes a tissues' tolerance for self-damage, which is required to protect against viral spread.

Increased random exploration in schizophrenia is associated with inflammation

One aspect of goal-directed behavior, which is known to be impaired in patients with schizophrenia (SZ), is balancing between exploiting a familiar choice with known reward value and exploring a lesser known, but potentially more rewarding option. Despite its relevance to several symptom domains of SZ, this has received little attention in SZ research. In addition, while there is increasing evidence that SZ is associated with chronic low-grade inflammation, few studies have investigated how this relates to specific behaviors, such as balancing exploration and exploitation. We therefore assessed behaviors underlying the exploration-exploitation trade-off using a three-armed bandit task in 45 patients with SZ and 19 healthy controls (HC). This task allowed us to dissociate goal-unrelated (random) from goal-related (directed) exploration and correlate them with psychopathological symptoms. Moreover, we assessed a broad range of inflammatory proteins in the blood and related them to bandit task behavior. We found that, compared to HC, patients with SZ showed reduced task performance. This impairment was due to a shift from exploitation to random exploration, which was associated with symptoms of disorganization. Relative to HC, patients with SZ showed a pro-inflammatory blood profile. Furthermore, high-sensitivity C-reactive protein (hsCRP) positively correlated with random exploration, but not with directed exploration or exploitation. In conclusion, we show that low-grade inflammation in patients with SZ is associated with random exploration, which can be considered a behavioral marker for disorganization. hsCRP may constitute a marker for severity of, and a potential treatment target for maladaptive exploratory behaviors.

Collective decision-making when quantity is more important than quality: Lessons from a kidnapping social parasite

Identifying the general principles that shape mechanisms of collective decision-making requires studies that span a diversity of ecological contexts. However, collective decision-making has only been explored in a handful of systems. Here, I investigate the ecologically mediated costs and benefits of collective decisions by socially parasitic kidnapping ants Temnothorax americanus over where to launch raids to steal host brood. I first investigate their sampling strategies and preferences with choice tests. Using more realistic spatial scales, I confirm the findings of others that colonies use a sequential choice strategy, and do not compare options simultaneously. I then ask which ecological conditions could favour the evolution of this strategy by testing the following hypotheses from optimal foraging and mate choice theories: (a) raiding decisions are time constrained or (b) search payoffs are low due to resource uniformity. Spatial distribution and phenological data on nest contents support the time constraints hypothesis. Host nests contain an optimal ratio of brood and workers for a brief period relative to discovery rates. Colonies therefore benefit from raiding most nests they find in this period rather than deliberating over the best choice, favouring host quantity over quality. The decision strategy for raids uncovered here contrasts with best-of-n collective decision-making found in other systems. These findings demonstrate that ecological constraints on information acquisition can alter how collectives process information.

Complex Relationship between Tunneling Patterns and Individual Behaviors in Termites

AbstractThe nests built by social insects are complex group-level structures that emerge from interactions among individuals following simple behavioral rules. Nest patterns vary among species, and the theory of complex systems predicts that there is no simple one-to-one relationship between variation in collective patterns and variation in individual behaviors. Therefore, a species-by-species comparison of the actual building process is essential to understand the mechanism producing diverse nest patterns. Here, we compare tunnel formation of three termite sp ecies and reveal two mechanisms producing interspecific variation: in one, a common behavioral rule yields distinct patterns via parameter tuning, and in the other, distinct rules produce similar patterns. We found that two related species transport sand in the same way using mandibles but build tunnels with different degrees of branching. The variation arises from different probabilities of choosing between two behavioral options at crowded tunnel faces: excavating the sidewall to make a new branch or waiting for clearance to extend the current tunnel. We further discovered that a third species independently evolved low-branched patterns using different building rules, namely, a bucket brigade that can excavate a crowded tunnel. Our findings emphasize the importance of direct comparative study of collective behaviors at both individual and group levels.

Modern termites inherited the potential of collective construction from their common ancestor

Animal collective behaviors give rise to various spatial patterns, such as the nests of social insects. These structures are built by individuals following a simple set of rules, slightly varying within and among species, to produce a large diversity of shapes. However, little is known about the origin and evolution of the behavioral mechanisms regulating nest structures. In this study, we discuss the perspective of inferring the evolution of collective behaviors behind pattern formations using a phylogenetic framework. We review the collective behaviors that can be described by a single set of behavioral rules, and for which variations of the environmental and behavioral parameter values produce diverse patterns. We propose that this mechanism could be at the origin of the pattern diversity observed among related species, and that, when they are placed in the proper conditions, species have the behavioral potential to form patterns observed in related species. The comparative analysis of shelter tube construction by lower termites is consistent with this hypothesis. Although the use of shelter tubes in natural conditions is variable among species, most modern species have the potential to build them, suggesting that the behavioral rules for shelter tube construction evolved once in the common ancestor of modern termites. Our study emphasizes that comparative studies of behavioral rules have the potential to shed light on the evolution of collective behaviors.

Interactions between worker ants may influence the growth of ant cemeteries

When an ant dies within a nest, a worker ant carries its corpse away from the nest and drops it onto a pile known as an ant cemetery. These ant cemeteries form cluster patterns, and the dynamics of the corpse piles have been studied experimentally. The aim of the present study was to investigate how sensitivity to the presence of nest-mates would influence the corpse-carrying behaviour of ants, and how this would impact the dynamics of corpse pile clustering. This was achieved by developing an agent-based computational model in which simulated ‘ants’ (the agents) carry and drop ‘corpses’, resulting in the growth of the corpse pile. In the model, the probability of an ant dropping a corpse was tuned according to the presence or absence of nest-mates. The pile dynamics of the resulting model showed a partial match with the time series evolution of corpse piles observed with real ants in previous experimental studies. Although the switch of probabilities is a thought experiment, our results suggest that the corpse-carrying behaviour of worker ants might be influenced by interactions with their nest-mates because there is evidence that ant behaviour can be influenced by encounter rates.

Colony size and initial conditions combine to shape colony reunification dynamics

Group cohesion and collective decision-making are important for many social animals, like social insects, whose societies depend on the coordinated action of individuals to complete collective tasks. A useful model for understanding collective, consensus-driven decision-making is the fluid nest selection dynamics of ant colonies. Certain ant species oscillate between occupying multiple nests simultaneously (polydomy) and reuniting at a single location (monodomy), but little is known about how colonies achieve a consensus around these dynamics. To investigate the factors underpinning the splitting-reunification dynamics of ants, we manipulated the availability and quality of nest sites for the ant Temnothorax rugatulus and measured the likelihood and speed of reunification from contrasting starting conditions. We found that pursuing reunification was more likely for smaller colonies, that rates of initial splitting were lower when colonies could coordinate their activity from a central hub, and that diluting colonies among additional sites did not impair reaching consensus on a single nest. We further found mixed support for a specific threshold of social density that prevents reunification (i.e., prolonged polydomy) and no evidence that nest quality influences reunification behavior. Together our data reveal that consensus driven decisions can be influenced by both external and intrinsic group-level factors and are in no way simple stereotyped processes.

Sources of intraspecific variation in the collective tempo and synchrony of ant societies

Populations of independently oscillating agents can sometimes synchronize. In the context of animal societies, conspicuous synchronization of activity is known in some social insects. However, the causes of variation in synchrony within and between species have received little attention. We repeatedly assessed the short-term activity cycle of ant colonies (Temnothorax rugatulus) and monitored the movements of individual workers and queens within nests. We detected persistent differences between colonies in the waveform properties of their collective activity oscillations, with some colonies consistently oscillating much more erratically than others. We further demonstrate that colony crowding reduces the rhythmicity (i.e., the consistent timing) of oscillations. Workers in both erratic and rhythmic colonies spend less time active than completely isolated workers, but workers in erratic colonies oscillate out of phase with one another. We further show that the queen's absence can impair the ability of colonies to synchronize worker activity and that behavioral differences between queens are linked with the waveform properties of their societies.

Symmetry breaking and pivotal individuals during the reunification of ant colonies

Maintenance of a social group requires the ability to reach consensus when faced with divisive choices. Thus, when migrating colonies of the ant Temnothorax rugatulus split among multiple sites, they can later reunify on the basis of queen location or differences in site quality. In this study, we found that colonies can reunify even without obvious cues to break the symmetry between sites. To learn how they do so, we observed both symmetric reunifications (between identical nests) and asymmetric reunifications (between nests of unequal quality) by colonies of individually marked ants. Both reunification types were accomplished by a tiny minority that carried nestmates from the 'losing' to the 'winning' site. Reunification effort was highly skewed in asymmetric splits, where the majority of the work was done by the first ant to transport, which nearly always came from the winning site. This contrasted with symmetric splits, where the initiator did not play an outsize role and was just as likely to come from the losing site. Symmetric reunifications were also characterized by high transporter attrition, which may help to prevent deadlocks. Tandem runs were abundant in both types and were typically led by transporters as they returned to the losing site to fetch another nestmate. Few tandem followers joined the transport effort, suggesting that tandem runs do not serve to recruit transporters but may have another, as yet unidentified role. Our results underscore the potentially large contribution of highly active individuals to group behaviour, even in decentralized societies such as ant colonies.

Different bees, different needs: how nest-site requirements have shaped the decision-making processes in homeless honeybees (Apis spp.)

During reproductive swarming, a honeybee swarm needs to decide on a new nest site and then move to the chosen site collectively. Most studies of swarming and nest-site selection are based on one species, Apis mellifera Natural colonies of A. mellifera live in tree cavities. The quality of the cavity is critical to the survival of a swarm. Other honeybee species nest in the open, and have less strict nest-site requirements, such as the open-nesting dwarf honeybee Apis floreaApis florea builds a nest comprised of a single comb suspended from a twig. For a cavity-nesting species, there is only a limited number of potential nest sites that can be located by a swarm, because suitable sites are scarce. By contrast, for an open-nesting species, there is an abundance of equally suitable twigs. While the decision-making process of cavity-nesting bees is geared towards selecting the best site possible, open-nesting species need to coordinate collective movement towards areas with potential nest sites. Here, we argue that the nest-site selection processes of A. florea and A. mellifera have been shaped by each species' specific nest-site requirements. Both species use the same behavioural algorithm, tuned to allow each species to solve their species-specific problem.This article is part of the theme issue 'Collective movement ecology'.

Phenotypic variability predicts decision accuracy in unicellular organisms

When deciding between different options, animals including humans face the dilemma that fast decisions tend to be erroneous, whereas accurate decisions tend to be relatively slow. Recently, it has been suggested that differences in the efficacy with which animals make a decision relate closely to individual behavioural differences. In this paper, we tested this hypothesis in a unique unicellular organism, the slime mould Physarum polycephalum. We first confirmed that slime moulds differed consistently in their exploratory behaviour from 'fast' to 'slow' explorers. Second, we showed that slow explorers made more accurate decisions than fast explorers. Third, we demonstrated that slime moulds integrated food cues in time and achieved higher accuracy when sampling time was longer. Lastly, we showed that in a competition context, fast explorers excelled when a single food source was offered, while slow explorers excelled when two food sources varying in quality were offered. Our results revealed that individual differences in accuracy were partly driven by differences in exploratory behaviour. These findings support the hypothesis that decision-making abilities are associated with behavioural types, even in unicellular organisms.

Social learning strategies regulate the wisdom and madness of interactive crowds

Why groups of individuals sometimes exhibit collective 'wisdom' and other times maladaptive 'herding' is an enduring conundrum. Here we show that this apparent conflict is regulated by the social learning strategies deployed. We examined the patterns of human social learning through an interactive online experiment with 699 participants, varying both task uncertainty and group size, then used hierarchical Bayesian model fitting to identify the individual learning strategies exhibited by participants. Challenging tasks elicit greater conformity among individuals, with rates of copying increasing with group size, leading to high probabilities of herding among large groups confronted with uncertainty. Conversely, the reduced social learning of small groups, and the greater probability that social information would be accurate for less-challenging tasks, generated 'wisdom of the crowd' effects in other circumstances. Our model-based approach provides evidence that the likelihood of collective intelligence versus herding can be predicted, resolving a long-standing puzzle in the literature.

Regulatory mechanism predates the evolution of self-organizing capacity in simulated ant-like robots

The evolution of complexity is one of the prime features of life on Earth. Although well accepted as the product of adaptation, the dynamics underlying the evolutionary build-up of complex adaptive systems remains poorly resolved. Using simulated robot swarms that exhibit ant-like group foraging with trail pheromones, we show that their self-organizing capacity paradoxically involves regulatory behavior that arises in advance. We focus on a traffic rule on their foraging trail as a regulatory trait. We allow the simulated robot swarms to evolve pheromone responsiveness and traffic rules simultaneously. In most cases, the traffic rule, initially arising as selectively neutral component behaviors, assists the group foraging system to bypass a fitness valley caused by overcrowding on the trail. Our study reveals a hitherto underappreciated role of regulatory mechanisms in the origin of complex adaptive systems, as well as highlights the importance of embodiment in the study of their evolution.

Parallel vs. comparative evaluation of alternative options by colonies and individuals of the ant Temnothorax rugatulus

Both a single ant and the colony to which it belongs can make decisions, but the underlying mechanisms may differ. Colonies are known to be less susceptible than lone ants to “choice overload”, whereby decision quality deteriorates with increasing number of options. We probed the basis of this difference, using the model system of nest-site selection by the ant Temnothorax rugatulus. We tested the applicability of two competing models originally developed to explain information-processing mechanisms in vertebrates. The Tug of War model states that concurrent alternatives are directly compared, so that choosing between two alternatives takes longer than accepting a single one. In contrast, the Sequential Choice Model assumes that options are examined in parallel, and action takes place once any option reaches a decision criterion, so that adding more options shortens time to act. We found that single ants matched the Tug of War model while colonies fitted the Sequential Choice model. Our study shows that algorithmic models for decision-making can serve to investigate vastly different domains, from vertebrate individuals to both individuals and colonies of social insects.

An agent-based model of nest-site selection in a mass-recruiting ant

Complex systems are modular entities which can collectively generate sophisticated emergent solutions through interactions based on simple, local rules. In this study, I use an agent-based model to elucidate how numerous individual-level components contribute to the collective decision process during house-hunting in a mass-recruiting ant species. Myrmecina nipponica combines the use of pheromone trails with a quorum decision rule in collective decisions among nest sites when searching for a new home. The model employed only individual-level rules but accurately emulated group-level properties observed in empirical studies. Simulations suggest that in this system i) both social and private information are necessary for effective decision making, ii) decision making was effective even with very low numbers of 'discriminating' individuals, iii) individual acceptance thresholds were more influential than quorum thresholds in tuning decisions to emphasise speed or accuracy, and iv) acceptance thresholds could also help tune decisions to suit environmental complexity. Similar findings in species using one-to-one recruitment suggest that some individual parameters, such as acceptance thresholds, may hold key functions in collective decision making regardless of the form of recruitment.

The Psychology of Superorganisms: Collective Decision Making by Insect Societies

Under the superorganism concept, insect societies are so tightly integrated that they possess features analogous to those of single organisms, including collective cognition. If so, colony function might fruitfully be studied using methods developed to understand individual animals. Here, we review research that uses psychological approaches to understand decision making by colonies. The application of neural models to collective choice shows fundamental similarities between how brains and colonies balance speed/accuracy trade-offs in decision making. Experimental analyses have explored collective rationality, cognitive capacity, and perceptual discrimination at both individual and colony levels. A major theme is the emergence of improved colony-level function from interactions among relatively less capable individuals. However, colonies also encounter performance costs due to their reliance on positive feedback, which generates consensus but can also amplify errors. Collective learning is a nascent field for the further application of psychological methods to colonies. The research strategy reviewed here shows how the superorganism concept can serve as more than an illustrative analogy.

A conceptual and computational framework for modelling and understanding the non-equilibrium gene regulatory networks of mouse embryonic stem cells

The capacity of pluripotent embryonic stem cells to differentiate into any cell type in the body makes them invaluable in the field of regenerative medicine. However, because of the complexity of both the core pluripotency network and the process of cell fate computation it is not yet possible to control the fate of stem cells. We present a theoretical model of stem cell fate computation that is based on Halley and Winkler’s Branching Process Theory (BPT) and on Greaves et al.’s agent-based computer simulation derived from that theoretical model. BPT abstracts the complex production and action of a Transcription Factor (TF) into a single critical branching process that may dissipate, maintain, or become supercritical. Here we take the single TF model and extend it to multiple interacting TFs, and build an agent-based simulation of multiple TFs to investigate the dynamics of such coupled systems. We have developed the simulation and the theoretical model together, in an iterative manner, with the aim of obtaining a deeper understanding of stem cell fate computation, in order to influence experimental efforts, which may in turn influence the outcome of cellular differentiation. The model used is an example of self-organization and could be more widely applicable to the modelling of other complex systems. The simulation based on this model, though currently limited in scope in terms of the biology it represents, supports the utility of the Halley and Winkler branching process model in describing the behaviour of stem cell gene regulatory networks. Our simulation demonstrates three key features: (i) the existence of a critical value of the branching process parameter, dependent on the details of the cistrome in question; (ii) the ability of an active cistrome to “ignite” an otherwise fully dissipated cistrome, and drive it to criticality; (iii) how coupling cistromes together can reduce their critical branching parameter values needed to drive them to criticality.

Pluripotent stem cells possess the capacity both to renew themselves indefinitely and to differentiate to any cell type in the body. Thus the ability to direct stem cell differentiation would have immense potential in regenerative medicine. There is a massive amount of biological data relevant to stem cells; here we exploit data relating to stem cell differentiation to help understand cell behaviour and complexity. These cells contain a dynamic, non-equilibrium network of genes regulated in part by transcription factors expressed by the network itself. Here we take an existing theoretical framework, Transcription Factor Branching Processes, which explains how these genetic networks can have critical behaviour, and can tip between low and full expression. We use this theory as the basis for the design and implementation of a computational simulation platform, which we then use to run a variety of simulation experiments, to gain a better understanding how these various transcription factors can combine, interact, and influence each other. The simulation parameters are derived from experimental data relating to the core factors in pluripotent stem cell differentiation. The simulation results determine the critical values of branching process parameters, and how these are modulated by the various interacting transcription factors.

Localized growth and branching random walks with time correlations

We generalize a model of growth over a disordered environment, to a large class of Itō processes. In particular, we study how the microscopic properties of the noise influence the macroscopic growth rate. The present model can account for growth processes in large dimensions and provides a bed to understand better the tradeoff between exploration and exploitation. An additional mapping to the Schrödinger equation readily provides a set of disorders for which this model can be solved exactly. This mean-field approach exhibits interesting features, such as a freezing transition and an optimal point of growth, which can be studied in detail, and gives yet another explanation for the occurrence of the Zipf law in complex, well-connected systems.

Model of the best-of-N nest-site selection process in honeybees

The ability of a honeybee swarm to select the best nest site plays a fundamental role in determining the future colony's fitness. To date, the nest-site selection process has mostly been modeled and theoretically analyzed for the case of binary decisions. However, when the number of alternative nests is larger than two, the decision-process dynamics qualitatively change. In this work, we extend previous analyses of a value-sensitive decision-making mechanism to a decision process among N nests. First, we present the decision-making dynamics in the symmetric case of N equal-quality nests. Then, we generalize our findings to a best-of-N decision scenario with one superior nest and N-1 inferior nests, previously studied empirically in bees and ants. Whereas previous binary models highlighted the crucial role of inhibitory stop-signaling, the key parameter in our new analysis is the relative time invested by swarm members in individual discovery and in signaling behaviors. Our new analysis reveals conflicting pressures on this ratio in symmetric and best-of-N decisions, which could be solved through a time-dependent signaling strategy. Additionally, our analysis suggests how ecological factors determining the density of suitable nest sites may have led to selective pressures for an optimal stable signaling ratio.

Response threshold variance as a basis of collective rationality

Determining the optimal choice among multiple options is necessary in various situations, and the collective rationality of groups has recently become a major topic of interest. Social insects are thought to make such optimal choices by collecting individuals' responses relating to an option's value (=a quality-graded response). However, this behaviour cannot explain the collective rationality of brains because neurons can make only 'yes/no' responses on the basis of the response threshold. Here, we elucidate the basic mechanism underlying the collective rationality of such simple units and show that an ant species uses this mechanism. A larger number of units respond 'yes' to the best option available to a collective decision-maker using only the yes/no mechanism; thus, the best option is always selected by majority decision. Colonies of the ant Myrmica kotokui preferred the better option in a binary choice experiment. The preference of a colony was demonstrated by the workers, which exhibited variable thresholds between two options' qualities. Our results demonstrate how a collective decision-maker comprising simple yes/no judgement units achieves collective rationality without using quality-graded responses. This mechanism has broad applicability to collective decision-making in brain neurons, swarm robotics and human societies.

Information Certainty Determines Social and Private Information Use in Ants

Decision-making in uncertain environments requires animals to evaluate, contrast and integrate various information sources to choose appropriate actions. In consensus-making groups, quorum responses are commonly used to combine private and social information, leading to both robust and flexible decisions. Here we show that in house-hunting ant colonies, individuals fine-tune the parameters of their quorum responses depending on their private knowledge: informed ants evaluating a familiar new nest rely relatively more on social than private information when the certainty of their private information is low, and vice versa. This indicates that the ants follow a highly sophisticated ‘copy-when-uncertain’ social learning strategy similar to that observed in a few vertebrate species. Using simulations, we further show that this strategy improves colony performance during emigrations and confers well-informed individuals more weight in the decision process, thus representing a novel mechanism for the emergence of leadership in collective decision-making.

A framework for modeling information propagation of biological systems at critical states

We explore the dynamics of information propagation at the critical state of a biologically inspired system by an individual-based computer model. "Quorum response", a type of social interaction which has been recognized taxonomically in animal groups, is applied as the sole interaction rule among individuals. In the model, we assume a truncated Gaussian distribution to depict the distribution of the individuals' vigilance level. Each individual can assume either a naïve state or an alarmed one and only switches from the former state to the latter one. If an individual has turned into an alarmed state, it stays in the state during the process of information propagation. Initially, each individual is set to be at the naïve state and information is tapped into the system by perturbing an individual at the boundaries (alerting it to the alarmed state). The system evolves as individuals turn into the alarmed state, according to the quorum response rules, consecutively. We find that by fine-tuning the parameters of the mean and the standard deviation of the Gaussian distribution, the system is poised at a critical state. We present the phase diagrams to exhibit that the parameter space is divided into a super-critical and a sub-critical zone, in which the dynamics of information propagation varies largely. We then investigate the effects of the individuals' mobility on the critical state, and allow a proportion of randomly chosen individuals to exchange their positions at each time step. We find that mobility breaks down criticality of the system.

Transcranial Stimulation over Frontopolar Cortex Elucidates the Choice Attributes and Neural Mechanisms Used to Resolve Exploration-Exploitation Trade-Offs

Optimal behavior requires striking a balance between exploiting tried-and-true options or exploring new possibilities. Neuroimaging studies have identified different brain regions in humans where neural activity is correlated with exploratory or exploitative behavior, but it is unclear whether this activity directly implements these choices or simply reflects a byproduct of the behavior. Moreover, it remains unknown whether arbitrating between exploration and exploitation can be influenced with exogenous methods, such as brain stimulation. In our study, we addressed these questions by selectively upregulating and downregulating neuronal excitability with anodal or cathodal transcranial direct current stimulation over right frontopolar cortex during a reward-learning task. This caused participants to make slower, more exploratory or faster, more exploitative decisions, respectively. Bayesian computational modeling revealed that stimulation affected how much participants took both expected and obtained rewards into account when choosing to exploit or explore: Cathodal stimulation resulted in an increased focus on the option expected to yield the highest payout, whereas anodal stimulation led to choices that were less influenced by anticipated payoff magnitudes and were more driven by recent negative reward prediction errors. These findings suggest that exploration is triggered by a neural mechanism that is sensitive to prior less-than-expected choice outcomes and thus pushes people to seek out alternative courses of action. Together, our findings establish a parsimonious neurobiological mechanism that causes exploration and exploitation, and they provide new insights into the choice features used by this mechanism to direct decision-making.

Interactions Increase Forager Availability and Activity in Harvester Ants

Social insect colonies use interactions among workers to regulate collective behavior. Harvester ant foragers interact in a chamber just inside the nest entrance, here called the 'entrance chamber'. Previous studies of the activation of foragers in red harvester ants show that an outgoing forager inside the nest experiences an increase in brief antennal contacts before it leaves the nest to forage. Here we compare the interaction rate experienced by foragers that left the nest and ants that did not. We found that ants in the entrance chamber that leave the nest to forage experienced more interactions than ants that descend to the deeper nest without foraging. Additionally, we found that the availability of foragers in the entrance chamber is associated with the rate of forager return. An increase in the rate of forager return leads to an increase in the rate at which ants descend to the deeper nest, which then stimulates more ants to ascend into the entrance chamber. Thus a higher rate of forager return leads to more available foragers in the entrance chamber. The highest density of interactions occurs near the nest entrance and the entrances of the tunnels from the entrance chamber to the deeper nest. Local interactions with returning foragers regulate both the activation of waiting foragers and the number of foragers available to be activated.

Polya's bees: A model of decentralized decision-making

How do social systems make decisions with no single individual in control? We observe that a variety of natural systems, including colonies of ants and bees and perhaps even neurons in the human brain, make decentralized decisions using common processes involving information search with positive feedback and consensus choice through quorum sensing. We model this process with an urn scheme that runs until hitting a threshold, and we characterize an inherent tradeoff between the speed and the accuracy of a decision. The proposed common mechanism provides a robust and effective means by which a decentralized system can navigate the speed-accuracy tradeoff and make reasonably good, quick decisions in a variety of environments. Additionally, consensus choice exhibits systemic risk aversion even while individuals are idiosyncratically risk-neutral. This too is adaptive. The model illustrates how natural systems make decentralized decisions, illuminating a mechanism that engineers of social and artificial systems could imitate.

Computational model of collective nest selection by ants with heterogeneous acceptance thresholds

Collective decision-making is a characteristic of societies ranging from ants to humans. The ant Temnothorax albipennis is known to use quorum sensing to collectively decide on a new home; emigration to a new nest site occurs when the number of ants favouring the new site becomes quorate. There are several possible mechanisms by which ant colonies can select the best nest site among alternatives based on a quorum mechanism. In this study, we use computational models to examine the implications of heterogeneous acceptance thresholds across individual ants in collective nest choice behaviour. We take a minimalist approach to develop a differential equation model and a corresponding non-spatial agent-based model. We show, consistent with existing empirical evidence, that heterogeneity in acceptance thresholds is a viable mechanism for efficient nest choice behaviour. In particular, we show that the proposed models show speed-accuracy trade-offs and speed-cohesion trade-offs when we vary the number of scouts or the quorum threshold.

Collective Movement in the Tibetan Macaques (Macaca thibetana): Early Joiners Write the Rule of the Game

Collective behavior has recently attracted a great deal of interest in both natural and social sciences. While the role of leadership has been closely scrutinized, the rules used by joiners in collective decision making have received far less attention. Two main hypotheses have been proposed concerning these rules: mimetism and quorum. Mimetism predicts that individuals are increasingly likely to join collective behavior as the number of participants increases. It can be further divided into selective mimetism, where relationships among the participants affect the process, and anonymous mimetism, where no such effect exists. Quorum predicts that a collective behavior occurs when the number of participants reaches a threshold. To probe into which rule is used in collective decision making, we conducted a study on the joining process in a group of free-ranging Tibetan macaques (Macaca thibetana) in Huangshan, China using a combination of all-occurrence and focal animal sampling methods. Our results show that the earlier individuals joined movements, the more central a role they occupied among the joining network. We also found that when less than three adults participated in the first five minutes of the joining process, no entire group movement occurred subsequently. When the number of these early joiners ranged from three to six, selective mimetism was used. This means higher rank or closer social affiliation of early joiners could be among the factors of deciding whether to participate in movements by group members. When the number of early joiners reached or exceeded seven, which was the simple majority of the group studied, entire group movement always occurred, meaning that the quorum rule was used. Putting together, Macaca thibetana used a combination of selective mimetism and quorum, and early joiners played a key role in deciding which rule should be used.

An extra dimension to decision-making in animals: the three-way trade-off between speed, effort per-unit-time and accuracy

The standard view in biology is that all animals, from bumblebees to human beings, face a trade-off between speed and accuracy as they search for resources and mates, and attempt to avoid predators. For example, the more time a forager spends out of cover gathering information about potential food sources the more likely it is to make accurate decisions about which sources are most rewarding. However, when the cost of time spent out of cover rises (e.g. in the presence of a predator) the optimal strategy is for the forager to spend less time gathering information and to accept a corresponding decline in the accuracy of its decisions. We suggest that this familiar picture is missing a crucial dimension: the amount of effort an animal expends on gathering information in each unit of time. This is important because an animal that can respond to changing time costs by modulating its level of effort per-unit-time does not have to accept the same decrease in accuracy that an animal limited to a simple speed-accuracy trade-off must bear in the same situation. Instead, it can direct additional effort towards (i) reducing the frequency of perceptual errors in the samples it gathers or (ii) increasing the number of samples it gathers per-unit-time. Both of these have the effect of allowing it to gather more accurate information within a given period of time. We use a modified version of a canonical model of decision-making (the sequential probability ratio test) to show that this ability to substitute effort for time confers a fitness advantage in the face of changing time costs. We predict that the ability to modulate effort levels will therefore be widespread in nature, and we lay out testable predictions that could be used to detect adaptive modulation of effort levels in laboratory and field studies. Our understanding of decision-making in all species, including our own, will be improved by this more ecologically-complete picture of the three-way tradeoff between time, effort per-unit-time and accuracy.

Efficient decision-making is vital to the lives of all animals, but the underlying principles of how they achieve this are not yet fully understood. Researchers studying decision-making have generally assumed that animals balance a two-way trade-off between speed and accuracy: the more time they spend gathering information, the more accurate their decisions will be, but the greater the cost they have to pay. We suggest that this picture is missing a crucial component: the effort that animals spend on gathering information within each unit of time. This is important because an animal that can change the amount of effort it invests per-unit-time can use this ability to maintain the accuracy of its decisions even when it reduces the amount of time it spends on them, and can therefore gain a fitness advantage. We predict that this ability to change effort levels should therefore be widespread in nature. This updated view of a three-way trade-off between speed, effort per-unit-time and accuracy will help behavioral ecologists, neuroscientists, economists and psychologists to understand decision-making better, and may also lead to the development of more efficient control algorithms for robot decision-makers.

Network-based identification of biomarkers coexpressed with multiple pathways

Unraveling complex molecular interactions and networks and incorporating clinical information in modeling will present a paradigm shift in molecular medicine. Embedding biological relevance via modeling molecular networks and pathways has become increasingly important for biomarker identification in cancer susceptibility and metastasis studies. Here, we give a comprehensive overview of computational methods used for biomarker identification, and provide a performance comparison of several network models used in studies of cancer susceptibility, disease progression, and prognostication. Specifically, we evaluated implication networks, Boolean networks, Bayesian networks, and Pearson’s correlation networks in constructing gene coexpression networks for identifying lung cancer diagnostic and prognostic biomarkers. The results show that implication networks, implemented in Genet package, identified sets of biomarkers that generated an accurate prediction of lung cancer risk and metastases; meanwhile, implication networks revealed more biologically relevant molecular interactions than Boolean networks, Bayesian networks, and Pearson’s correlation networks when evaluated with MSigDB database.

Sensory optimization by stochastic tuning

Individually, visual neurons are each selective for several aspects of stimulation, such as stimulus location, frequency content, and speed. Collectively, the neurons implement the visual system's preferential sensitivity to some stimuli over others, manifested in behavioral sensitivity functions. We ask how the individual neurons are coordinated to optimize visual sensitivity. We model synaptic plasticity in a generic neural circuit and find that stochastic changes in strengths of synaptic connections entail fluctuations in parameters of neural receptive fields. The fluctuations correlate with uncertainty of sensory measurement in individual neurons: The higher the uncertainty the larger the amplitude of fluctuation. We show that this simple relationship is sufficient for the stochastic fluctuations to steer sensitivities of neurons toward a characteristic distribution, from which follows a sensitivity function observed in human psychophysics and which is predicted by a theory of optimal allocation of receptive fields. The optimal allocation arises in our simulations without supervision or feedback about system performance and independently of coupling between neurons, making the system highly adaptive and sensitive to prevailing stimulation.

Individual variation in exploratory behaviour improves speed and accuracy of collective nest selection by Argentine ants

Collective behaviours are influenced by the behavioural composition of the group. For example, a collective behaviour may emerge from the average behaviour of the group's constituents, or be driven by a few key individuals that catalyse the behaviour of others in the group. When ant colonies collectively relocate to a new nest site, there is an inherent trade-off between the speed and accuracy of their decision of where to move due to the time it takes to gather information. Thus, variation among workers in exploratory behaviour, which allows gathering information about potential new nest sites, may impact the ability of a colony to move quickly into a suitable new nest. The invasive Argentine ant, Linepithema humile, expands its range locally through the dispersal and establishment of propagules: groups of ants and queens. We examine whether the success of these groups in rapidly finding a suitable nest site is affected by their behavioural composition. We compared nest choice speed and accuracy among groups of all-exploratory, all-nonexploratory and half-exploratory-half-nonexploratory individuals. We show that exploratory individuals improve both the speed and accuracy of collective nest choice, and that exploratory individuals have additive, not synergistic, effects on nest site selection. By integrating an examination of behaviour into the study of invasive species we shed light on the mechanisms that impact the progression of invasion.

Collective decision-making in microbes

Microbes are intensely social organisms that routinely cooperate and coordinate their activities to express elaborate population level phenotypes. Such coordination requires a process of collective decision-making, in which individuals detect and collate information not only from their physical environment, but also from their social environment, in order to arrive at an appropriately calibrated response. Here, we present a conceptual overview of collective decision-making as it applies to all group-living organisms; we introduce key concepts and principles developed in the context of animal and human group decisions; and we discuss, with appropriate examples, the applicability of each of these concepts in microbial contexts. In particular, we discuss the roles of information pooling, control skew, speed vs. accuracy trade-offs, local feedbacks, quorum thresholds, conflicts of interest, and the reliability of social information. We conclude that collective decision-making in microbes shares many features with collective decision-making in higher taxa, and we call for greater integration between this fledgling field and other allied areas of research, including in the humanities and the physical sciences.

A context-dependent alarm signal in the ant Temnothorax rugatulus

Because collective cognition emerges from local signaling among group members, deciphering communication systems is critical to understanding underlying mechanisms. Alarm signals are widespread in the social insects and can elicit a variety of behavioral responses to danger, but the functional plasticity of these signals has not been well studied. Here we report an alarm pheromone in the ant Temnothorax rugatulus that elicits two different behaviors depending on context. When an ant was tethered inside an unfamiliar nest site and unable to move freely, she released a pheromone from her mandibular gland that signaled other ants to reject this nest as a potential new home, presumably to avoid potential danger. When the same pheromone was presented near the ants' home nest, they were instead attracted to it, presumably to respond to a threat to the colony. We used coupled gas chromatography/mass spectrometry to identify candidate compounds from the mandibular gland and tested each one in a nest choice bioassay. We found that 2,5-dimethylpyrazine was sufficient to induce rejection of a marked new nest and also to attract ants when released at the home nest. This is the first detailed investigation of chemical communication in the leptothoracine ants. We discuss the possibility that this pheromone's deterrent function can improve an emigrating colony's nest site selection performance.

A mechanism for value-sensitive decision-making

We present a dynamical systems analysis of a decision-making mechanism inspired by collective choice in house-hunting honeybee swarms, revealing the crucial role of cross-inhibitory 'stop-signalling' in improving the decision-making capabilities. We show that strength of cross-inhibition is a decision-parameter influencing how decisions depend both on the difference in value and on the mean value of the alternatives; this is in contrast to many previous mechanistic models of decision-making, which are typically sensitive to decision accuracy rather than the value of the option chosen. The strength of cross-inhibition determines when deadlock over similarly valued alternatives is maintained or broken, as a function of the mean value; thus, changes in cross-inhibition strength allow adaptive time-dependent decision-making strategies. Cross-inhibition also tunes the minimum difference between alternatives required for reliable discrimination, in a manner similar to Weber's law of just-noticeable difference. Finally, cross-inhibition tunes the speed-accuracy trade-off realised when differences in the values of the alternatives are sufficiently large to matter. We propose that the model, and the significant role of the values of the alternatives, may describe other decision-making systems, including intracellular regulatory circuits, and simple neural circuits, and may provide guidance in the design of decision-making algorithms for artificial systems, particularly those functioning without centralised control.

Harvester ant colony variation in foraging activity and response to humidity

Collective behavior is produced by interactions among individuals. Differences among groups in individual response to interactions can lead to ecologically important variation among groups in collective behavior. Here we examine variation among colonies in the foraging behavior of the harvester ant, Pogonomyrmex barbatus. Previous work shows how colonies regulate foraging in response to food availability and desiccation costs: the rate at which outgoing foragers leave the nest depends on the rate at which foragers return with food. To examine how colonies vary in response to humidity and in foraging rate, we performed field experiments that manipulated forager return rate in 94 trials with 17 colonies over 3 years. We found that the effect of returning foragers on the rate of outgoing foragers increases with humidity. There are consistent differences among colonies in foraging activity that persist from year to year.