Search bioRxiv⌕ Search

Biology subjects

Romanczuk, P.

Publications and source records attributed to Romanczuk, P..

17 recordsLinked to original sources
bioRxiv2026

Tracking human foragers and their prey reveals adaptive predator-prey dynamics

Hunting for mobile prey is thought to have played a key role in hominin evolution, by providing high-quality nutrition that supported the development of the exceptionally large human brain. However, human-prey dynamics remain poorly understood because studies have not yet tracked human foragers and their prey simultaneously. Here, we employ high resolution tracking of groups of human foragers (ice-fishers) and their prey (fish shoals) to study human-prey dynamics. Our results show that foragers adaptively combined personal and social information in deciding where to forage and for how long, closely matching the prey distribution. Prey responded dynamically to human exploitation, showing increased attraction to fishing activity, alongside decreased biting probability. Furthermore, we found that foragers adaptively relied on memory, preferentially returning to areas with high prey presence, particularly when their current return rate was low. Our results show how human foragers overcome the challenges of extracting invisible, mobile and reactive prey by tightly tuning patch-selection, patch-leaving and patch-return decisions to the distribution and behaviour of their prey.

animal behavior and cognition↗
bioRxiv2025

Deterministic dynamics of distributional multi-agent reinforcement learning

Understanding how cognition shapes behavior across contexts remains a fundamental challenge for many disciplines. In particular, for the optimism heuristic-i.e., the tendency to overweight positive (relative to negative) information-knowledge remains fragmented, with models developed in specific domains in isolation. Here, we present a unifying computational framework by deriving the deterministic dynamics of distributional multi-agent reinforcement learning. Our approach discretizes reward distributions through a finite set of neurons, consistent with recent empirical findings on distributional coding in the brain. We validate our framework by reproducing established results across three iconic domains spanning individual exploration versus exploitation, social coordination, and risky choice. Beyond validation, we uncover novel interactions among optimism, reward discretization, and temporal discounting. Specifically, we identify conditions under which choice hysteresis and path-dependent strategies emerge, suggesting that perseveration results from neural reward discretization rather than constituting an independent heuristic. We further reveal "individual dilemmas": circumstances where agents gravitate toward suboptimal yet stable strategies, offering a mechanistic explanation for incoherent choice patterns. Our framework bridges neuroscience, psychology, and collective behavior, enabling empirically testable hypotheses about how cognitive biases propagate from individual cognition to social outcomes in complex environments.

neuroscience↗
bioRxiv2025

Social Information Quality Determines Learned Collective Foraging Strategies

Collective foraging is widespread across the animal kingdom, allowing animals to more effectively discover resources. However, collective foragers need to balance a key trade off between private exploration and using social information. Social information can come in very distinct forms, ranging from simple positional cues to complex payoff information. However, how the types of available social cues and environmental volatility shape collective foraging behavior is not well understood. We address this using a spatially-explicit model in which agents track a mobile resource via multi-agent reinforcement learning. Agents choose between random exploration, private tracking, and social attraction. We systematically varied resource volatility and the type of available social cues to analyze their effect on individual and collective behavior. Our results show that the quality of social information dictates the emerging collective behavior. Low-quality social cues (e.g., positions, actions) result in a fragile strategy that is effective in stable environments but fails as volatility increases. Conversely, high-quality social information (e.g., payoffs) enables behavioral diversity: Agents selectively copy others and flexibly change between individual tracking or exploration depending on the environmental volatility. Our findings identify the interplay between information quality and ecological context as an important mechanism governing the emergence of distinct forms of collective behavior from individual decision rules.

biophysics↗
bioRxiv2025

Real-Time Human Interaction with Virtual Swarms in Shared Physical Space

Human-swarm interaction (HSI) explores how humans engage with distributed collective systems, aiming to incorporate human cognition into scalable and robust robotic swarms. While most HSI research focuses on remote teleoperation via engineered interfaces, real-world integration of swarms into everyday tasks requires natural, embodied interactions in shared physical spaces. To address the limitations of traditional teleoperation studies, and the high resource demands of using physical robot swarms for HSI research, we introduce CoBe XR, a spatial augmented reality system that projects virtual swarms into the physical environment of the human operator. CoBe XR enables real-time, fine-grained, natural interaction between humans and swarm-like agents through full-body movement without dedicated control interfaces or prior training. As a proof-of-concept, we present a behavioral study involving 40 participants who influenced swarm behavior solely through walking. Our results show that human participants were able to adapt to the collective dynamics of the swarm and control it through natural perception-motion control in a shared physical space. We argue that similar extended reality systems can not only reveal how humans perceive and adapt to collective dynamics, but they offer a general platform to understand human behavior or an intermediate solution to design embodied robot swarms.

animal behavior and cognition↗
bioRxiv2024

Better and faster collective decisions by larger fish shoals in the wild

Studies on collective cognition have provided many examples of decision-making benefits in terms of animals sharing information about predators, prey or resources in their environment. It has been shown how the efficient spread of adaptive information within groups can provide benefits which increase with group size. Little is known, however, to which extent groups also amplify maladaptive information such as false alarms and whether such costs reduce or even nullify the above benefits. Here, we investigated fish shoals in the wild that responded collec-tively with escape dives when attacked by birds. We analysed the response of shoals in reac-tion to hard-to-detect bird attacks and similar but harmless flybys as a function of shoal size. With increasing shoal size fish increasingly detected predator attacks (true positives) while their false alarms remained constant. Therefore, larger shoals became better at correctly clas-sifying potentially dangerous stimuli rather than becoming more sensitive to all stimuli poten-tially related to attacks. In addition, decision time decreased with increasing shoal size. Larger shoals were thus able to mitigate two major trade-offs inherent in solitary decision making: the trade-off between true and false positives and the trade-off between speed and accuracy. We report performance increases at shoal sizes of tens of thousands of fish and pose chal-lenges for the modelling of the underlying mechanisms.

animal behavior and cognition↗
bioRxiv2024

Collective anti-predator escape manoeuvres through optimal attack and avoidance strategies

The collective dynamics of self-organised systems emerge from the decision rules agents use to respond to each other and to external forces. This is evident in groups of animals under attack from predators, where understanding collective escape patterns requires evaluating the risks and rewards associated with particular social rules, prey escape behaviour, and predator attack strategies. Here, we find that the emergence of the fountain effect, a common collective pattern observed when animal groups evade predators, is the outcome of rules designed to maximise individual survival chances given predator hunting decisions. Using drone-based empirical observations of schooling sardine prey (Sardinops sagax caerulea) attacked by striped marlin (Kajikia audax), we first find the majority of attacks produce fountain effects, with the dynamics of these escapes dependent on the predators attack direction. Then, using a spatially-explicit agent-based model of predator-prey dynamics, we show that fountain manoeuvres can emerge from combining an optimal individual prey escape angle with social interactions. The escape rule appears to prioritise maximising the distance to the predator and creates conflict in the effectiveness of predators attacks and the preys avoidance, explaining the empirically observed predators attack strategies and the fountain evasions produced by prey. Overall, we identify the proximate and ultimate explanations for fountain effects and more generally highlight that the collective patterns of self-organised predatory-prey systems can be understood by considering both social escape rules and attack strategies.

ecology↗
bioRxiv2023

Visual social information use in collective foraging

Collective dynamics emerge from individual-level decisions, yet we still poorly understand the link between individual-level decision-making processes and collective outcomes in realistic physical systems. Using collective foraging to study the key trade-off between personal and social information use, we present a mechanistic, spatially-explicit agent-based model that combines individual-level evidence accumulation of personal and (visual) social cues with particle-based movement. Under idealized conditions without physical constraints, our mechanistic framework reproduces findings from established probabilistic models, but explains how individual-level decision processes generate collective outcomes in a bottom-up way. In clustered environments, groups performed best if agents reacted strongly to social information, while in uniform environments, individualistic search was most beneficial. Incorporating different real-world physical and perceptual constraints profoundly shaped collective performance, and could even buffer maladaptive herding by facilitating self-organized exploration. Our study uncovers the mechanisms linking individual cognition to collective outcomes in human and animal foraging and paves the way for decentralized robotic applications. Significance statementFinding and collecting rewards in heterogeneous environments is key for adaptive collective behavior in humans, animals and machines. We present an open agent-based simulation framework to study how social information use shapes collective foraging from the bottom up. Our model combines individual evidence accumulation with spatially explicit movement. Our results connect individual-level decisions to collective dynamics in realistic physical environments, highlighting the key role of real-world constraints, thereby bringing us closer to embodied collective intelligence. Our work introduces a flexible platform to study the interplay between individual cognitive and perceptual biases, agents physical environment and the resulting collective dynamics and thus paves the way for fully decentralized mobile robot applications.

animal behavior and cognition↗
bioRxiv2023

Moderate confirmation bias enhances collective decision-making

Humans tend to give more weight to information confirming their beliefs than to information that disconfirms them. Nevertheless, this apparent irrationality has been shown to improve individual decision-making under uncertainty. However, little is known about this bias impact on collective decision-making. Here, we investigate the conditions under which confirmation bias is beneficial or detrimental to collective decision-making. To do so, we develop a Collective Asymmetric Reinforcement Learning (CARL) model in which artificial agents observe others actions and rewards, and update this information asymmetrically. We use agent-based simulations to study how confirmation bias affects collective performance on a two-armed bandit task, and how resource scarcity, group size and bias strength modulate this effect. We find that a confirmation bias benefits group learning across a wide range of resource-scarcity conditions. Moreover, we discover that, past a critical bias strength, resource abundance favors the emergence of two different performance regimes, one of which is suboptimal. In addition, we find that this regime bifurcation comes with polarization in small groups of agents. Overall, our results suggest the existence of an optimal, moderate level of confirmation bias for collective decision-making. AUTHOR SUMMARYWhen we give more weight to information that confirms our existing beliefs, it typically has a negative impact on learning and decision-making. However, our study shows that moderate confirmation bias can actually improve collective decision-making when multiple reinforcement learning agents learn together in a social context. This finding has important implications for policymakers who engage in fighting against societal polarization and the spreading of misinformation. It can also inspire the development of artificial, distributed learning algorithms. Based on our research, we recommend not directly targeting confirmation bias but instead focusing on its underlying factors, such as group size, individual incentives, and the interactions between bias and the environment (such as filter bubbles).

animal behavior and cognition↗
bioRxiv2022

Experienced social partners hinder learning performance in naive clonal fish

Social learning is widely assumed to enhance individual learning efficiency, particularly when naive observers have access to skilled demonstrators, yet the conditions under which this assumption holds remain poorly understood. Here, we investigated how individual learning is influenced by the skill level of social partners. We predicted that naive individuals would benefit from observing experienced conspecifics, yet found the opposite: trained partners significantly impaired naive individuals learning performance, while trained individuals remained unaffected by their partners skill level. We conducted experiments in near-identical individuals, using the all-female clonal Amazon molly (Poecilia formosa) to test whether these fish can learn an operant foraging task, whether individuals differ consistently in learning ability, and whether partner skill level influences learning performance. Using an operant conditioning paradigm over five days, half of the fish were trained to locate food inside a cylinder, whereas the remaining individuals received food randomly dispersed within their tank. Trained individuals subsequently visited the cylinder more frequently than randomly fed individuals and exhibited consistent individual differences in learning performance. In a second phase, fish were allowed to observe a conspecific while individual training was either continued (for trained individuals) or initiated (for naive individuals). We found that trained individuals did not benefit from the presence of a partner, regardless of the partners proficiency, but consistently outperformed naive individuals. In contrast, naive individuals showed reduced learning performance when paired with experienced partners, but not with naive partners. Together, our results indicate that Amazon mollies achieve this foraging task through individual learning and exhibit stable differences in learning ability. Moreover, social learning depends on both the learners own skill level and that of the social partner, such that observing an experienced conspecific may, in some cases, impede rather than enhance learning.

animal behavior and cognition↗
bioRxiv2020

Group-level patterns emerge from individual speed as revealed by an extremely social robotic fish

Understanding the emergence of collective behaviour has long been a key research focus in the natural sciences. Besides the fundamental role of social interaction rules, a combination of theoretical and empirical work indicates individual speed may be a key process that drives the collective behaviour of animal groups. Socially-induced changes in speed by interacting animals make it difficult to isolate the effects of individual speed on group-level behaviours. Here we tackled this issue by pairing guppies with a biomimetic robot. We used a closed-loop tracking and feedback system to let a robotic fish naturally interact with a live partner in real time, and programmed it to strongly copy and follow its partner’s movements while lacking any preferred movement speed or directionality of its own. We show that individual differences in guppies’ movement speed were highly repeatable and shaped key collective patterns: higher individual speeds resulted in stronger leadership, lower cohesion, higher alignment, and better temporal coordination in the pairs. By combining the strengths of individual-based models and observational work with state-of-the-art robotics, we provide novel evidence that individual speed is a key, fundamental process in the emergence of collective behaviour.Competing Interest StatementThe authors have declared no competing interest.View Full Text

animal behavior and cognition↗
bioRxiv2020

The wisdom of stalemates: consensus and clustering as filtering mechanisms for improving collective accuracy

Groups of organisms, from bacteria to fish schools to human societies, depend on their ability to make accurate decisions in an uncertain world. Most models of collective decision-making assume that groups reach a consensus during a decision-making bout, often through simple majority rule. In many natural and sociological systems, however, groups may fail to reach consensus, resulting in stalemates. Here, we build on opinion dynamics and collective wisdom models to examine how stalemates may affect the wisdom of crowds. For simple environments, where individuals have access to independent sources of information, we find that stalemates improve collective accuracy by selectively filtering out incorrect decisions. In complex environments, where individuals have access to both shared and independent information, this effect is even more pronounced, restoring the wisdom of crowds in regions of parameter space where large groups perform poorly when making decisions using majority rule. We identify network properties that tune the system between consensus and accuracy, providing mechanisms by which animals, or evolution, could dynamically adjust the collective decision-making process in response to the reward structure of the possible outcomes. Overall, these results highlight the adaptive potential of stalemale filtering for improving the decision-making abilities of group-living animals.

animal behavior and cognition↗
bioRxiv2019

Harvesting-induced evolution of collectivebehavior in a fish

Intensive and size-selective harvesting is an evolutionary driver of life-history as well as individual behavioral traits. Yet, whether and to what degree harvesting modifies the collective behavior of exploited species is largely unknown. We present a multi-generation harvest selection experiment with zebrafish (Danio rerio) as a model species to understand the effects of size-selective harvesting on shoaling behavior. The experimental system is based on a large-harvested (typical of most wild capture fisheries targeting larger size classes) and small-harvested (typical of specialized fisheries and gape-limited predators targeting smaller size classes) selection lines. By combining high resolution tracking of fish behavior with computational agent-based modeling we show that shoal cohesion changed in the direction expected by a trade-off between vigilance and the use of social cues. In particular, we document a decrease of vigilance in the small-harvested line, which was linked to an increase in the attention to social cues, favoring more cohesive shoals. Opposing outcomes were found for the large-harvested line, which formed less cohesive shoals. Using the agent-based model we outline possible consequences of changes is shoaling behavior for both fishing and natural mortality. The changes in shoaling induced by large size-selective harvesting may decrease fishing mortality, but increase mortality by natural predators. Our work suggests an insofar overlooked evolutionary mechanism by which size-selective harvesting can affect mortality and in turn population dynamics of exploited fish.

ecology↗
bioRxiv2019

A model of collective behavior based purely on vision

Classical models of collective behavior often take a \"birds-eye perspective,\" assuming that individuals have access to social information that is not directly available (e.g., the behavior of individuals outside of their field of view). Despite the explanatory success of those models, it is now thought that a better understanding needs to incorporate of the perception of the individual, i.e. how internal and external information are acquired and processed. In particular, vision has appeared to be a central feature to gather external information and influence the collective organization of the group. Here we show that a vision based model of collective behavior is sufficient to generate organized collective behavior in the absence of spatial representation and collision. Our work suggests a novel approach for development of purely vision-based autonomous swarm robotic systems, and formulates a mathematical framework for exploration of perception-based interactions and how they differ from physical ones. Thus, it is of broader relevance for self-organization in complex systems, neuroscience, behavioral sciences and engineering.

animal behavior and cognition↗
bioRxiv2018

Social networks in the presence and absence of visual cues

We compared the social dynamics of two populations of the live-bearing Atlantic molly (Poecilia mexicana) that live in adjacent habitats with very different predator regimes: cave mollies that inhabit a low-predation environment inside a sulfidic cave with a low density of predatory water bugs (Belostoma sp.), and mollies that live directly outside the cave (henceforth called \"surface\" mollies) in a high-predation environment with a high density of fish-eating birds. We filmed the social interactions of marked fish in both environments and analysed their social network dynamics using a Markov model under two different fish densities of 12 and 6 fish per 0.36 m2. As expected, surface mollies spent overall much more time social than cave mollies. This difference in overall social time was a result of surface mollies being less likely to discontinue social contact (once they had a social partner) and being more likely to resume social contact (once alone) than cave mollies. Interestingly surface mollies were also less likely to leave a current social partner than cave mollies. At low density, mollies (in both environments) were expected to show reduced social encounters which should dramatically change their social dynamics. Surface mollies, however, displayed an ability to maintain their social dynamics at low density (primarily by reducing the convex polygon spanned by the group) which was not observed in cave mollies. Despite the fact that we only compared two populations, our data provide a mechanistic explanation for density compensations of social dynamics that have also been observed in other fish species and give an example of how comparisons between the social dynamics of different populations can be made that go beyond conventional network analyses.

animal behavior and cognition↗
bioRxiv2018

Searching for structure in collective systems

Collective systems such as fish schools, bird flocks, and neural networks are comprised of many mutually-influencing individuals, often without long-term leaders, well-defined hierarchies, or persistent relationships. The remarkably organized group-level behaviors readily observable in these systems contrast with the ad hoc, often difficult to observe, and complex interactions among their constituents. While these complex individual-level dynamics are ultimately the drivers of group-level coordination, they do not necessarily offer the most parsimonious description of a groups macroscopic properties. Rather, the factors underlying group organization may be better described at some intermediate, mesoscopic scale. We introduce a novel method from information-theoretic first principles to find a compressed description of a system based on the actions and mutual dependencies of its constituents, thus revealing the natural structure of the collective. We emphasize that this method is computationally tractable and requires neither pairwise nor Gaussian assumptions about individual interactions.

animal behavior and cognition↗
bioRxiv2018

Guppies prefer to follow large (robot) leaders irrespective of own size

Body size is often assumed to determine how successful an individual can lead others with larger individuals being more likely to lead than smaller ones. However, direct evidence for such a relation is scarce. Furthermore, even if larger individuals are more likely to lead, body size correlates often with specific behavioral patterns (e.g., swimming capacity) and it is thus unclear whether larger individuals are more often followed than smaller ones because they are larger or because they behave in a certain way. To control for behavioral differences among differentially-sized leaders, we used biomimetic robotic fish - Robofish - of different sizes. Robofish is accepted as a conspecific by live guppies (Poecilia reticulata) and provides standardized behaviors irrespective of its size. We specifically asked whether larger leaders are preferentially followed when behavior is controlled for and whether the preferences of followers depend on their own body size or their risk taking behavior ( boldness). We found that live guppies followed larger Robofish leaders closer than smaller ones and this pattern was independent of the followers own body size as well as risk-taking behavior. This is the first study that shows a bigger is better pattern in leadership in shoaling fish that is fully independent of behavioral differences between differentially-sized leaders and followers own size and personality.

animal behavior and cognition↗
bioRxiv2018

Using a robotic fish to investigate individual differences in social responsiveness in the guppy

Responding towards the actions of others is one of the most important behavioral traits whenever animals of the same species interact. Mutual influences among interacting individuals may modulate the social responsiveness seen and thus makes it often difficult to study the level and variation of individuality in responsiveness. Here, biomimetic robots (BRs) that are accepted as conspecifics but controlled by the experimenter can be a useful tool. Studying the interactions of live animals with BRs allows pinpointing the live animals level of responsiveness by removing confounding mutuality. In this paper, we show that live guppies (Poecilia reticulata) exhibit consistent differences among each other in their responsiveness when interacting with a biomimetic fish robot - Robofish - and a live companion. It has been repeatedly suggested that social responsiveness correlates with other individual behavioral traits like risk-taking behavior ( boldness) or activity level. We tested this assumption in a second experiment. Interestingly, our detailed analysis of individual differences in social responsiveness using the Robofish, suggests that responsiveness is an independent trait, not part of a larger behavioral syndrome formed by boldness and activity.

animal behavior and cognition↗