Neural Bases of Collective Social Behavior and Group Interactions across Species

Abstract

Collective group behavior is a key form of social interaction observed across animal species, including humans. These collective behaviors, such as flocking and herding, emerge from individual actions within complex group dynamics and can often lead to group phenomena such as social norms and trends. By investigating different species—insects, fish, birds, and mammals—ongoing efforts have revealed both shared and species-specific neural processes involved in collective social dynamics. Newly evolving technologies, such as motion tracking and behavioral sequencing, have enabled the detailed measurement of group behaviors, helping to elucidate the neural mechanisms behind cooperation, coordination, and competition. Advancements in wireless neuronal recordings and neuromodulation have also provided new insights into brain circuits involved in group behaviors and how manipulation of agents such as dopamine, serotonin, and oxytocin shapes social dynamics. Moreover, emerging evidence has suggested that certain neurons may be selectively tuned to specific features of collective behavior, such as group movement, identity, or social rank, pointing to the existence of neural coding strategies necessary for representing group dynamics. Based on this growing knowledge, we hypothesize that distributed neural circuits spanning regions like the prefrontal cortex, medial temporal lobe, and sensory areas interact to support collective social cognition. We also discuss how interdisciplinary approaches that integrate neuroscience, psychology, sociology, neurolinguistics, electrical engineering, and computer science can improve our understanding of group dynamics and propose future directions for studying collective social behavior.

Significance Statement

Understanding how individual actions give rise to collective group behavior—and how collective dynamics, in turn, shape individual behavior—is central to multiple research areas, including neuroscience. This study integrates insights from animal models, neuroendocrinology, and emerging technologies to explore the neural, hormonal, and computational mechanisms underlying collective behavior across species. By examining distributed neural circuits, social cues, and the influence of neuromodulators alongside analytical models, the work highlights both shared and species-specific strategies of group interaction. The future framework developed from these findings not only advances our understanding of collective social cognition but also informs future applications in mental health and artificial systems designed to emulate social intelligence.

Introduction

Living in groups offers many benefits such as access to shared resources, safety, and learning opportunities, many of which contribute to survival and adaptability (McGlynn, 2010). When individuals come together as a group, their interactions often generate emergent patterns and behaviors that cannot be explained solely by the actions of any one member. Collective group behaviors are also characterized by their flexibility and responsiveness to changing conditions, providing a key basis of how individuals adapt to and are dynamically influenced by the behaviors of others. As a complex phenomenon shaped by environmental and social factors, such as group characteristics, social hierarchies, and environmental contexts, collective behavior plays a crucial role in social interactions across many species, including humans.

The study of collective group behavior has gathered attention across various fields, from zoology, biology, and ecology (Couzin, 2009; Ioannou and Laskowski, 2023; Toyokawa, 2023) to human sociology (da Costa et al., 2023; González, 2024) and artificial intelligence (Werfel et al., 2014; Alizadehsani et al., 2023). In neuroscience, the study of collective behavior is particularly important as it sheds light on the brain mechanisms underlying social interaction, decision-making, and group patterns. By examining how individuals contribute to and respond to the combined collective dynamics of others, we can understand the neural circuits and pathways involved in complex social behaviors. This research is inherently interdisciplinary, integrating perspectives from neuroscience, psychology, sociology, neurolinguistics, electrical engineering, and computer science and aims to explore how individuals act collectively within groups. It also spans various animal models, including insects, fish, birds, bats, rodents, nonhuman primates (NHPs), and humans which are commonly employed to study these questions. Each model provides unique advantages for examining specific neural processes implicated in social interaction and collective decision-making, together allowing researchers to identify both species-specific and generalizable mechanisms.

Here, we provide a comprehensive overview of the neuroscience of collective group behavior, beginning with foundational concepts and theories related to group and collective dynamics. We provide essential background information and highlight key animal models commonly used in the field, along with their featured advantages. We also highlight new analytical approaches and discuss how technological advances have enhanced our ability to study the neural and behavioral dynamics in collective social behavior across various species as well as how such findings may extend to human studies. We then explore the neural mechanisms underlying these behaviors, including neurophysiology, neuromodulation, and neuroendocrine factors that may influence social interactions and their coordination within groups (Fig. 1).

Figure 1.

A defining characteristic of collective behavior. Collective behavior is often modulated by external factors, such as environmental cues, threats, or rewards, which may further influence group dynamics. Internally, the underlying neural mechanisms, encompassing neurophysiology, neuromodulation, and neuroendocrine factors, play a critical role in shaping social interactions and their coordination within groups.

Finally, we propose central questions on how the neural mechanisms that govern dyadic (pairwise) interactions compare and translate to those involved in larger-scale collective behaviors. We also consider the broader implications of such findings and suggest directions for future research. In particular, we hypothesize that collective social behavior relies not only on distinct brain regions that mediate social behavior [e.g., the amygdala and prefrontal cortex (PFC)] or their circuits but also on the dynamic integration of multiple neural systems including those that may uniquely subserve group decisions. By integrating knowledge from diverse species and disciplines, this review aims to enhance our understanding of the brain mechanisms that drive collective group behavior and point to future directions that could be taken.

The study of collective group behavior

Collective group behavior refers to the ways individuals within social groups interact and coordinate their actions, leading to emergent group-level behaviors. These behaviors are typically governed by simple rules that, when followed by individuals, produce complex group patterns. A key feature of collective behavior is that the group exhibits properties that are not present at the individual level and emerge from local interactions, such as alignment of movement or information sharing during swarming or consensus decision-making. These behaviors are also often influenced by external factors such as environmental cues, threats, or rewards that further shape the group's dynamics (Fig. 1). Further embedded within this framework is the paradox of individuality (Parrish and Edelstein-Keshet, 1999), which underscores the tension between individual-level strategies and group-level outcomes, emphasizing how self-interested behaviors can often either enhance or undermine the cohesion, stability, and functionality of the group dynamics. This paradox raises important questions about the extent to which individuality can be separated from the social context in which it operates, as well as the neural mechanisms that underlie this dynamic interplay.

Collective behaviors are adaptive, allowing animals to function efficiently within social groups and respond dynamically to environmental challenges. Examples of collective responses can be observed across biological systems and range from the development of social norms, such as cultural fads and fashions in humans (Bikhchandani et al., 1992), to group dynamics, including coordinated movements, social foraging, and cooperative hunting (Creel and Creel, 1995; Couzin, 2006; Krivoruchko et al., 2024) in many animal species. Localized conformity and the fragility of mass behaviors can often be explained by informational cascades, where individuals mimic the actions of their predecessors (Bikhchandani et al., 1992). Groups also often rely on collective sensing mechanisms to respond to environmental gradients (Berdahl et al., 2013). Studies have shown that social animals exhibit behaviors influenced by interactive effects such as alignment, attraction, and repulsion, as seen in the shoaling of fish (Pitcher, 1986; Katz et al., 2011). Moreover, conformity and social pressure are critical factors often seen in humans and NHPs, whereby individuals adjust their behaviors to align with what is believed to be expected by others in the group (Asch, 1955; Whiten et al., 2005). Taken together, these social dynamics highlight the dual role of individual cognition (e.g., personal decision-making) and group-level processes (e.g., emergent group interactions) in shaping the collective behavior of groups. Studying the distinctions between individual and collective behavior is, therefore, a key tool for understanding the transition from individual decision-making to group-level processes and how group behavior enhances the fitness and survival of individuals (Couzin et al., 2002).

To understand how collective behaviors emerge, it is essential to define the relationship between social context and information processing within groups. The dynamic between group members influences how information is transmitted and can affect the group's collective responses to external influences such as chemical alarm cues (Cao et al., 2023) or predator threats. For example, the rapid transmission of information is crucial for creating an effective system that enables efficient responses to predators (Papadopoulou et al., 2022) such as in sticklebacks who exhibit increased synchronization that reduces a predator's ability to target individual prey (Bode et al., 2010). This type of collective behavior is also a remarkable example model for multiagent systems that have informed our understanding of both computational and real-world phenomena (Urbaniak et al., 2024). These collective phenomena therefore have significant implications, not only for understanding cognitive neuroscience and animal behavior but also for addressing challenges in artificial intelligence with multiagent systems (Cui and Yasseri, 2024). Here, we describe ongoing progress in our understanding of group behavior, the paired dynamics that influence group interactions, early investigations into the neural mechanisms that underlie collective group patterns, and new technologies that could expand our ability to study group dynamics.

Studying collective group behavior across animal species

Animal models have provided an invaluable opportunity to study collective group behavior, offering insights into the internal (e.g., biological), external (e.g., environmental), and social factors that drive such behaviors. A variety of species, including insects, fish, birds, mammals, NHPs, and humans have been used to examine different aspects of collective group dynamics and their underlying neural bases. Comprehensively, these animal models enable both controlled laboratory experiments and field observations, revealing how individual actions aggregate to produce group-level behaviors. They also provide a framework for understanding the shared mechanisms that drive group interactions across different biological systems.

Insects, like locusts and honeybees, for example, serve as ideal models for studying swarm behaviors and the influence of environmental factors, such as heat stress, on collective motion and decision-making (Ariel and Ayali, 2015; Jhawar et al., 2023; Traniello and Avarguès-Weber, 2023; Tarpy, 2024). They also offer unique insight into how groups allocate effort under uncertain conditions, particularly when individuals rely solely on local information and coordination emerges through distributed interactions and collective motion (Carlesso et al., 2023; Sayin et al., 2025). Furthermore, their relatively simple neural and behavioral systems, when compared with mammals, also make them well suited for observation and experimentation in controlled laboratory environments. Additionally, their behaviors are easily manipulated and studied using genetic and neural perturbation techniques (Buhl and Rogers, 2016; Feinerman and Korman, 2017).

Birds, such as storks, on the other hand, are often useful for examining collective navigation and migration patterns. These studies shed light on how animals coordinate long-distance movements and how leadership emerges within groups (Flack et al., 2018). Birds also often live in large groups and engage in complex vocal communication, which facilitates rich social interactions (Guggenberger et al., 2022). Moreover, they exhibit distinct behaviors, such as flocking in flight, cooperative breeding, and group foraging, highlighting their adaptive nature and their ability to efficiently coordinate actions (Cavagna et al., 2015; Suzuki et al., 2016; Cohen et al., 2020; Farine, 2022).

Fish models, such as golden shiners and zebrafish, are particularly useful for studying collective movement and decision-making in medium- to large-sized groups. These models help elucidate how individuals synchronize their behavior to form cohesive schools and respond to external factors (Wang et al., 2022). They provide valuable features for analyzing how environmental factors, such as shifts in preferred habitat (Kleinhappel et al., 2023) and hydrodynamic indicators (Liu et al., 2023) as well as sensory inputs like visual cues (Miles et al., 2024), influence group behavior. Moreover, the group size significantly influences the observed behaviors of groups, particularly in how visual information affects collective detection (Davidson et al., 2021), therefore allowing the scales of these group effects to be studied in detail.

The study of mammals, including mice, rats, and bats, can offer significant insights into the neurophysiological mechanisms underlying social behavior (Shemesh et al., 2013; Lee and Noh, 2015; Rose et al., 2021; Forli and Yartsev, 2023). Their tractability in laboratory settings enhances their utility in exploring the neural basis of social and collective group behaviors. They also often display consistent and reproducible social behaviors across individuals, making them a common model organism for genetic, hormonal, and chemical manipulation studies. For example, mice and rats have been widely used to examine how specific brain regions contribute to group dynamics and coordinated social interactions (Williamson et al., 2016; Li et al., 2022; Kong et al., 2023; Zilkha et al., 2023). Furthermore, bat colonies display highly flexible, context-dependent interactions and cooperative behaviors, characterized by dynamic fission–fusion group structures that facilitate the study of complex social cognition (Kerth, 2008; Kerth et al., 2011; Ray et al., 2025). Especially, the specific vocalizations of bats also offer opportunities to study the dynamics of social communication (Prat et al., 2015; Rose et al., 2021).

NHPs, such as rhesus monkeys and marmosets, display many forms of collective behavior similar to those observed in humans, including cooperation, competition, and conflict resolution (Bernstein and Mason, 1963; Maestripieri and Hoffman, 2011; Miller et al., 2016). Their strong sense of group identity enables researchers to study specific social interactions with greater precision. Their study (Báez-Mendoza et al., 2021) also provides the opportunity to understand how individual members represent and contribute to interactions within the group. Recent findings further reveal that dominant male macaques with greater facial expressiveness tend to be more socially connected and maintain more cohesive social groups (Whitehouse et al., 2024). Additionally, their behavioral and neurobiological basis of vocal communication within group also highlights their broader significance as research models (Eliades and Miller, 2017). Leveraging such features could offer valuable opportunities for advancing our understanding of collective behavior in primates, since their brain structure and function closely resemble those of humans, making them a popular choice for studies on collective social behavior. Moreover, small groups of NHPs can be studied in both natural and laboratory environments, providing diverse opportunities to examine their social dynamics (Tyree et al., 2023; Franch et al., 2024; Testard et al., 2024).

Finally, humans, though more challenging to study in controlled laboratory settings, exhibit highly complex collective behaviors across diverse platforms, ranging from social media to political and economic systems. Examples include stock market dynamics (Maskawa, 2016; Alves et al., 2020) and the emergence of fashion trends (Seiler and Rumpel, 2023), which are often driven by the collective actions of groups. In addition, humans are capable of forming intricate social structures, norms, and institutions that shape large-scale interactions (Hinde, 1976; Kessinger et al., 2023; Kish Bar-On and Lamm, 2023; Noyes et al., 2023). They are also well suited for both observational and experimental research through methods such as behavioral tracking (Jenkins et al., 2018; Shruthi et al., 2022), surveys (Du et al., 2023), and analysis of digital footprints (Albuquerque et al., 2023). Most notably, humans use language to communicate with other individuals, coordinate actions, and be able to drive complex collective behaviors (Jamali et al., 2021, 2024; Wang et al., 2025). These features offer unique insights into processes such as cultural evolution, decision-making, and large-scale social influence. This linguistic and cognitive capacity surpasses that of other animal species and offers a unique opportunity to study the mechanisms by which humans interact with others. Moreover, research involving humans holds significant potential for directly informing and implementing targeted interventions.

In Figure 2, we summarized the key characteristics of each animal research model discussed in this section. Here, the models highlighted were selected based on their relevance and representation in the existing literature. Many of the advantages attributed to these models are not necessarily unique to a single species or exclusive to specific taxonomic groups. For instance, various animals—such as rhesus monkeys, mice, and bats—display pronounced dominance hierarchies that influence their interactions and allow for detailed study of social identity (Báez-Mendoza et al., 2021; Kong et al., 2023; Ray et al., 2025) and status (Li et al., 2022; Lee et al., 2024). Furthermore, as we will explore in the next sections, advancements in technology and analytical methods are helping to bridge gaps among research models, enabling broader questions and hypotheses to be tested across species.

Figure 2.

An overview of various animal models used in the study of collective group behavior, highlighting their advantages to understanding different aspects of group dynamics.

Experimental and analytical models for collective group behavior

Different analytical models have been developed to understand the underlying dynamics of groups, offering crucial initial insights into these complex phenomena and how they may relate to human behavior. Early mathematical frameworks, such as the Selfish Herd Model by Hamilton (1971), for example, highlighted the role of geometrical probabilities in group formation. More recent models have also integrated environmental data and social interaction metrics, as demonstrated in fish schooling (Couzin et al., 2002; Davidson et al., 2021), which emphasize detection efficiency and spatial dynamics. Another approach is swarm intelligence, which is based on decentralized decision-making, where individual agents follow simple rules that lead to complex group behaviors. Moreover, O'Bryan et al. (2020) explored how swarm behavior can enhance our understanding of collective intelligence in human teams, which could, in turn, have practical applications for understanding of team performance and the development of comparative frameworks for improving group-level dynamics.

We schematically illustrate the integration of spatial dynamics and behavioral cues in the study of collective group behavior in Figure 3. Prior to developing such analytical models, it is essential to consider how to measure subjects’ behavior from a wide range of perspectives. Research often tracks subjects’ spatial locations in a two-dimensional (2D) plane (Shemesh et al., 2013; Berman et al., 2014; Lee and Noh, 2015; Tavares et al., 2015; Williamson et al., 2016; Davidson et al., 2021; Li et al., 2022; Wang et al., 2022; Zilkha et al., 2023; Sayin et al., 2025) and sometimes extends to a three-dimensional (3D) plane (Couzin et al., 2002; Flack et al., 2018; Forli and Yartsev, 2023; Testard et al., 2024). This broad viewpoint enables the analysis of both individual and group dynamics. To zoom in on these perspectives, we must also be able to examine behavioral patterns at the individual level. The transition from 2D to 3D space modeling has allowed researchers to investigate more intricate behaviors, such as changes in postures, spatial positioning, and nuanced interactions between individuals within groups (Nourizonoz et al., 2020; Dunn et al., 2021; Lin et al., 2024; Menegas et al., 2024; Klibaite et al., 2025). It has also broadened the scope of analysis to encompass nonmotion-related behaviors, including viewing patterns (Capozzi et al., 2019; Bastien and Romanczuk, 2020; Maran et al., 2020; Wu et al., 2024), vocalization features (Prat et al., 2015; Logue and Krupp, 2016; Suzuki et al., 2016; Eliades and Miller, 2017; Sagot et al., 2018; Cohen et al., 2020; Rose et al., 2021; Guggenberger et al., 2022; Demartsev et al., 2023; Xie et al., 2024), and language-based communication in humans (Maass et al., 1989; Van Swol and Kane, 2019; Ashokkumar and Pennebaker, 2022; von Heiseler, 2022; Bishop and Lerch, 2023). Furthermore, leveraging computer vision techniques, we can detect body parts and associate them with the subject's specific pose or gesture dynamics.

Figure 3.

An illustration of analytical frameworks for investigating collective group behavior. Top left: Spatial locations and trajectories of individuals can be tracked in both 2D and 3D environments to identify group movement patterns. Top right: A range of social cues, including motion, vocalization, gaze patterns, and language, can be captured across different species. Bottom: These multimodal data streams, encompassing spatial dynamics, auditory signals, visual perspectives, behavioral sequences, neural activity patterns, and more, can be integrated and analyzed using analytical models such as dimensionality reduction (e.g., t-SNE) and deep learning model (e.g., large language models).

The recent development of motion-tracking tools has significantly advanced the precision and depth of behavioral analysis (Aung et al., 2023). Animal pose estimation and tracking algorithms such as DeepLapCut (Mathis et al., 2018) and Social LEAP Estimates Animal Poses (Pereira et al., 2022), for example, have enabled researchers to study the detailed behaviors of individuals within groups. These individual-level motions can then be extended to understand behavior at the group level. Furthermore, tools such as motion sequencing (Wiltschko et al., 2020), Simple Behavioral Analysis (Goodwin et al., 2024), and A-SOiD (Tillmann et al., 2024) have allowed for high-resolution tracking and recognition of behaviors which has greatly enhanced the capacity to quantify group dynamics in naturalistic settings. Additionally, automated tracking technologies like EthoLoop (Nourizonoz et al., 2020), BeesBook (Wario et al., 2017; Boenisch et al., 2018), and multiplexed videography systems (Ebbesen and Froemke, 2022) have enhanced the ability to recognize complex social dynamics. These behavioral data streams can be integrated with multimodal inputs, such as spatial dynamics, auditory signals, visual perspectives, neural activity patterns, and more. Combined, they can be analyzed using computational techniques, including dimensionality reduction (e.g., t-SNE) and deep learning approaches (e.g., large language models) to extract meaningful patterns (Johnson et al., 2020; Li et al., 2022; Suzuki and Arita, 2024).

Lastly, there has been an increasing drive to incorporate computational modeling with neuroimaging to link social computations with underlying brain mechanisms (Lockwood and Klein-Flügge, 2021; Cowley et al., 2024; Wittmann et al., 2025). A variety of analytical approaches have also been extended to gain a better understanding of the relationship between individual- and group-level decision–making (Beal Cohen et al., 2021; Garg et al., 2024). For example, Mezey et al. (2024) introduced a mechanistic, spatially explicit agent-based model that integrates individual evidence accumulation based on personal and visual social cues with particle-based movements. Agent-based models provide a powerful framework for simulating group behavior while incorporating individual differences, social influences, and environmental context. Importantly, these approaches offer a powerful framework for contextualizing the role of neural substrates across species by bridging the gap between individual-level neural mechanisms and emergent group-level dynamics. By implementing simple, biologically plausible rules at the neural or behavioral level, these models can generate complex patterns of collective behavior. These approaches also allow researchers to systematically explore how factors such as the group size, individual traits, and interaction rules contribute to collective outcomes, including cohesion, coordination, and information transfer.

Taken together, these tools have helped facilitate the precise monitoring of individual movements and behaviors and the ability to analyze complex group data (Costa-Pereira et al., 2022). By capturing higher-resolution data, these tools could provide deeper insights into the mechanisms underlying collective behavior and have the potential of greatly improving our understanding of group-level coordination and decision-making.

Neural mechanisms underlying group behaviors

Building on early investigations into paired interactions, a growing number of neurophysiological studies have begun to identify some of the brain mechanisms that may support collective group behavior. They have also begun to identify neurons that encode aspects of group interactions and that can track the complex collective dynamic of individuals within a group. In primates, for example, neurons have been identified in the dorsomedial PFC (dmPFC) that encode both the agency and identity of others, facilitating agent-specific interactions (Báez-Mendoza et al., 2021). Similarly, in humans, the dmPFC has been shown to reflect social preferences and their alignments with those of other group members (Izuma and Adolphs, 2013). Furthermore, fMRI research has demonstrated multiple causal roles of the human ventromedial PFC in prosocial behavior and its motivation (Lockwood et al., 2024). Neurons in the dorsolateral PFC of NHPs have been shown to encode social decisions as they relate to cooperation (Franch et al., 2024). Neurons in the dmPFC of mice also adaptively represent the social rank of the animals in relation to others (Li et al., 2022), together suggesting that the PFC may serve as a key hub for processing and controlling both paired and group interactions (Fan et al., 2024).

In addition to prefrontal regions, parts of the medial temporal lobe have been shown to encode information about others and to provide computations necessary for effective group interactions (Edmonds et al., 2024). The amygdala, for example, does not process social status in isolation but instead appears to evaluate social interactions based on environmental and situational cues (Lee et al., 2024) and to reflect identity-specific defensive behaviors (Kim et al., 2020). Hippocampal cells in primates, on the other hand, have been shown to reflect the social relevance of the attended object (Staudigl et al., 2022) and to integrate multisensory input that represents the group network and social identity of others (Tyree et al., 2023). The activities of hippocampal cells in bats have also been shown to encode social information such as identity, sex, hierarchy, and affiliation within a group (Ray et al., 2025). They have also been shown to track spatial locations under naturalistic conditions (Eliav et al., 2021) and to support spatially structured group behaviors (Forli and Yartsev, 2023). Furthermore, overlapping functioning has been observed in the human hippocampus (Tavares et al., 2015; Quian Quiroga, 2023; Attaallah et al., 2024), together revealing neural processes that may support the ability of individuals to interact within social groups.

Cross-species comparisons are particularly valuable for identifying shared features and species-specific adaptations. For instance, one critical component of collective behavior is communication, which plays a central role in coordinating group actions. For example, prior studies have explored the role of the locus ceruleus, the caudomedial nidopallium, and the premotor nucleus high vocal center in birds (Cohen et al., 2020; Katic et al., 2022) and the frontal cortex in bats during group vocal interactions (Rose et al., 2021). In primates, studies have examined the connectivity between the PFC and PMC to the auditory cortex circuit of marmosets (Eliades and Miller, 2017), as well as the involvement of the human PFC in language processing (Jamali et al., 2024; Khanna et al., 2024). These comparisons can help expand our understanding of how different species rely on specific brain regions and circuits to facilitate communication within groups and raise important questions about the evolutionary convergence and divergence of these mechanisms.

Finally, the advancement of neural recording systems has significantly enhanced both the flexibility of experimental design and the quality and quantity of collected neurophysiological data. Wireless neural recording systems have been developed to reliably record and transmit neural activity to external receivers in real time (Lee et al., 2021; Kwon et al., 2024), allowing researchers to record neural data from ad libitum moving subjects. This technology enables real-time neural monitoring, making it especially suitable for situations that require live feedback. Furthermore, customized wireless systems have facilitated simultaneous recordings from multiple interacting individuals, allowing researchers to analyze the group social behavior (Kim et al., 2020; Shin et al., 2022). Neural loggers, on the other hand, are devices designed to record neural data locally onto onboard memory or external storage, such as flash memory. The advent of neural loggers has therefore enabled the study of naturalistic behaviors across large areas, including 3D flight, underwater swimming, and translocation in the wild. These tools have also provided valuable insights into the neural underpinnings of social behavior in natural environments (Ide and Takahashi, 2022) as well as social interactions in unrestrained conditions (Testard et al., 2024). For example, Rose et al. (2021) demonstrated how intra- and inter-brain activity patterns in groups of bats are modulated by social context and individual social preferences when communicating. Neural loggers have also enabled researchers to examine human subjects with chronically implanted electrodes as they perform real-world spatial navigation and observation tasks (Stangl et al., 2021) further highlighting the potential impact of these technologies in understanding human social behavior.

Taken together, these emerging approaches have begun to reveal the intricate interplay between individual-level neural processes, paired interactions, and group-level behavior, offering pathways for interdisciplinary research into collective cognition and behavior. Future directions will likely focus on examining larger group dynamics and simultaneous recordings from multiple individuals as they perform structured and naturalistic tasks. They may also include the implementation of neuromodulatory techniques such as that used by Topalovic et al. (2023), through Neuro-stack, which enables bidirectional closed-loop neuromodulation and recording of neural data during both stationary and ambulatory behaviors. Finally, future investigations may focus on expanding our understanding of how stress (Smoulder et al., 2024) and social influence (Izuma and Adolphs, 2013) affect group behavior as well as how social constructs such as group conformity, social exclusion, and decision-making under pressure lead to adaptive and maladaptive group behavior.

Neuroendocrine and neurochemical influences on collective behavior

Another important question in social neuroscience is how the neuroendocrine system affects social behavior. Neuroendocrine and neurochemical mechanisms play a pivotal role in modulating social interactions, including paired and group interactions (Jolles et al., 2020). These mechanisms, for example, often influence emotional and motivational states that can, in aggregate, powerfully shape group behavior. The precise relationship between neuroendocrine systems and group-level interactions, however, is still poorly understood.

Several animal models have been instrumental in exploring neuroendocrine differences across taxa. The observable behaviors of social insect populations such as mating rituals, foraging, and feeding behaviors, for instance, have been shown to be influenced by over 150 unique neuropeptides (Robinson et al., 2019). Similarly, adrenal glucocorticoids as well as gonadal androgens and estrogens widely influence the regulation of aggression, territoriality, and social competition in mammalian species. They have also been shown to influence social communication and signaling and to underlie sexually dimorphic variations in social communication (Freiler and Smith, 2023). In some primates, individual differences in endocrine response and function have been further associated with dominance rank (Sapolsky and Share, 1994). Endocrine variations have also been shown to influence group cooperation, aggression, mating behaviors, dominance, and bonding. These behaviors can differ based on dominance hierarchies and, in humans, have been shown to influence trust and group leadership (Derntl et al., 2014; van der Meij et al., 2016).

Dopamine is another dominant neurochemical that has been shown to influence a range of social behaviors, from aggression and affiliation to bonding, parenting, and mating. Rodent studies, for example, indicate that reward experiences motivate the underlying search for affiliative social stimuli and that this function can be traced to dopaminergic neurons in the VTA and nucleus accumbens (Liu et al., 2010). In the context of group behaviors, the effects of dopamine have also been shown to differ based on the social dominance hierarchy of individuals in relation to their conspecifics (Maestripieri, 2010). Dopamine also plays a key part in social reward motivation, especially for behaviors related to status obtainment and mating (Campbell et al., 2010). Serotonin, on the other hand, appears to be negatively related to aggression in primates, such that low central nervous system serotonin function is a predictor of strongly violent, aggressive tendencies (Kiser et al., 2012). In macaques, this also includes greater impulsivity and a greater propensity for risk-taking, although it is unclear whether serotonin itself is related to aggression in all contexts (Maestripieri, 2010).

Furthermore, hormones like oxytocin and vasopressin are known to play significant roles in promoting social bonding and cooperation, which are essential for collective behavior in many species. For example, in many species of rodents and primates, oxytocin has been linked to behavior-mediated increases in affiliative behaviors, social recognition, and—in humans—group cohesion (Insel, 2010; Feldman, 2012; Anacker and Beery, 2013; Spengler et al., 2017; Yao et al., 2017; Nair and Chang, 2022). In fish and birds, changes in corticosterone levels in response to stressors can affect group behavior, influencing individual differences in sociality and reactivity to stressors (Cockrem, 2007). In NHPs, findings that oxytocin modulates social attention, and its association with prosocial behavior, social tendencies, and dyadic relationships, further underscore its role in shaping complex social dynamics (Chang and Platt, 2014; Putnam et al., 2016; Jiang and Platt, 2018; Festante et al., 2021; Liu et al., 2025). Meanwhile, across species, oxytocin continues to influence sociosexual behaviors, aggression, and involvement in between-group conflict through homologues, including isotocin (affecting social behavior in fish) and mesotocin (in reptile and bird species, as well as lungfish, amphibians, and many marsupial species; Triki et al., 2022).

Therefore, despite preservation across taxa, the respective influence of the neuroendocrine system on group behaviors varies widely across species and can lead to a broad range of social behaviors. These differences may relate to variation in social complexity, group structure, and adaptive challenges in species such as primates, birds, and insects. Many neuropeptide hormones, however, indicate the preservation of related structures and behavioral functions, and therefore studying their function is essential to our understanding of collective group behavior. Additionally, neuroendocrine mechanisms are critical for modulating individual roles within a group. For instance, the formation of dominance hierarchies in animal groups may be influenced by levels of testosterone, especially in cases of male–male aggression, while stress responses triggered by cortisol can lead to a variety of changes in social behavior and group interaction patterns, especially in cases of potential or realized status loss (Sherman and Mehta, 2020; Milewski et al., 2022). Together, these variations highlight the diversity and complexity by which neurotransmitters and the neuroendocrine system affect social behavior across animals and emphasize the need to study collective group behavior in different species.

Discussion

The study of collective behavior integrates insights from behavioral studies, computational models, and neuroscience in order to build a robust framework for understanding complex social systems. From foundational principles like conformity and emergent dynamics to the neural substrates of social interaction, this multidisciplinary field continues to uncover critical mechanisms that govern collective action. Collective behavior plays a pivotal role in survival and has been broadly shaped by evolution across species. Without the benefit of distributed feedback, groups risk making suboptimal or harmful decisions, as individuals may lack the ability to fully detect and evaluate the consequences of their actions. Collective intelligence mitigates this issue by leveraging diverse feedback, which is particularly effective in larger groups where input tends to be more varied and informative. Through social interaction, individuals are able to gain information from others and increase the efficiency and accuracy of decision-making both individually and collectively.

Future research should aim to investigate how distributed feedback mechanisms operate across various species and contexts. Pagan et al. (2025), for example, proposed that neural signatures of behavior can predict broader context-dependent computations, indicating that specific neural populations may flexibly encode social information depending on environmental demands. Building on this framework, a deeper understanding of these processes can enrich the study of collective behavior by revealing how individuals not only influence their group's dynamics but also how group-level feedback shapes individuals’ behavior. Exploring this bidirectional relationship will provide valuable insight into the mechanisms driving collective decision-making and adaptive behaviors across species. These investigations could also have practical applications, such as improving group coordination strategies, enhancing collaborative decision-making, and predicting collective outcomes.

One major limitation in studying collective behavior through animal models might be the difficulty in capturing higher-order cognitive processes. While species like rodents, fish, and birds exhibit sophisticated group behaviors, these are often governed by relatively simple, reflexive rules such as local alignment or “attraction-repulsion” dynamics. As a result, such models may not adequately reflect complex social functions like perspective-taking, social norm compliance, and strategic decision-making. Though useful for identifying basic principles of collective coordination, they provide a limited window into the cognitive mechanisms underlying more abstract and deliberative forms of group behavior. Overcoming these limitations may require using cognitively advanced species or enhancing experimental designs with advanced technologies. For example, NHPs are capable of redirecting the visual attention of their social partners, inducing gaze-following saccades (Emery, 2000; Shepherd, 2010). This behavior suggests an intrinsic interest in monitoring the exchange of agonistic and affiliative social information within their group (Champ et al., 2022), with social gaze interaction being processed extensively in the PFC–amygdala circuitry (Dal Monte et al., 2022). Moreover, recent advances in noninvasive eye-tracking and retinal view reconstruction in fish have begun to reveal how individuals’ visual perspectives contribute to large-scale collective behavior (Wu et al., 2024). Building on these findings, future research could further explore the causal neural mechanisms underlying social gaze coordination and collective behavior by integrating real-time neural recording and manipulation techniques in ad libitum interacting animal groups.

Studying human behavior also offers unique avenues for exploration, particularly regarding the role of language. Humans possess the distinct ability to communicate through structured speech, for example, which broadly affects collective decision-making and group dynamics. Recent studies have also begun to identify some of the neuronal elements associated with language in the human PFC (Jamali et al., 2024; Khanna et al., 2024) which could open the door to further understanding social communication and interaction. These findings also highlight the importance of studying human-unique abilities and their effects on group behavior. Such insights may further bridge the gap between behavioral neuroscience and human societal dynamics, opening new frontiers for interdisciplinary research. Additionally, the field of noninvasive neuromodulation is rapidly advancing, with acoustic techniques like ultrasound offering the ability to transiently modulate neural activity (Blackmore et al., 2019). These techniques are being applied to human subjects via transcranial ultrasound stimulation (Attali et al., 2023) and are currently being explored in clinical trials, including recent work targeting major depressive disorder (Oh et al., 2024). Incorporating such methods into collective behavioral research could yield valuable insights into the neural mechanisms that underlie group interactions and their dysfunction.

Beyond studying the neural circuits and mechanisms that are involved in group behavior, the study of neuroendocrinology in the context of collective behavior is also essential to fully understanding how hormones and neurotransmitters influence social behavior, decision-making, and group dynamics. The activity of neuromodulators may also influence the many instinctive—or otherwise rapid—social behaviors underlying collective behavior through the communication of signaling molecules, including steroid hormones (i.e., testosterone, estradiol), neuropeptides (i.e., oxytocin, vasopressin), and neurotransmitters (i.e., serotonin, dopamine), with a variety of target neurons through volume transmission and/or synaptic transmission. Estradiol, in particular, has been found to rapidly influence aggression in two species (Peromyscus mice and song sparrows) through membrane-initiated effects. Furthermore, distinguishing between nuclear-initiated (genomic) and membrane-initiated (nongenomic) mechanisms of steroid hormone action in future research may help to further clarify the form and extent of their social behavioral effects (Heimovics et al., 2015). Understanding these mechanisms can offer a clearer perspective of the changes caused by interactions involving mood, status, and social structure. For instance, while hormones like adrenal glucocorticoids are responsive to individual positions and changes in group structure (Cavigelli and Caruso, 2015), they have not yet been consistently linked to specific patterns or functions of social systems or shifts in collective behavior. Further study is also needed to understand how specific neuroendocrine mechanisms preserve their functions across various social behavioral contexts, ultimately leading to observable differences in group dynamics and collective actions.

A promising future direction in collective group behavior research involves moving beyond the study of single brain regions to explore the coordinated activity across distributed neural circuits. Recent findings, for instance, suggest that population-level neural representations contribute to social behavior by mapping model units to visual neurons that can reveal population codes underlying social behavior (Cowley et al., 2024). Higher-order visual areas—previously shown to encode visual flow and movement (Castro et al., 2024)—may uniquely participate in representing the movements of multiple individuals, such as the flocking of birds or the schooling of fish, thereby linking behavior to neural computation within sensory systems. In addition to this, the orbitofrontal cortex (OFC) has also recently emerged as a key node in social cognition, particularly in facilitating cooperative behavior and social decision-making (Shi et al., 2023). This function may be especially important for tracking the behavior of multiple individuals (Báez-Mendoza et al., 2021). Future studies should aim to test hypotheses on how the PFC and OFC interact with other regions, such as the medial temporal lobe and sensory regions, during collective group tasks. Employing approaches like multiregion neural recordings, computational modeling, and casual perturbation techniques will be crucial for uncovering the distributed neural computations that govern complex group dynamics. This network-level perspective has the potential to significantly deepen our understanding of the neural architecture underlying collective social cognition.

Finally, to gain a better understanding of collective group behavior, there has been a growing effort to track the complex collective dynamics of group interactions. At the same time, there is a need to better understand and track the behavior of animals, including humans, in naturalistic settings. For example, novel systems have recently been developed that utilize drones to collect animal behavioral data in the wild, paired with 3D spatial modeling and computer vision algorithms for analysis (Koger et al., 2023). Advancements in mobile neuroimaging technology have also enabled the simultaneous recording of diverse behavioral features and neural data in human research (Stangl et al., 2023). Integrating these innovative approaches may make it possible to design field experiments that collect both behavioral and neural data simultaneously. This synergy holds significant promise for advancing our understanding of brain function across various domains, including the dynamics of collective behavior in nature. Many species, such as insects and birds, display complex emergent behaviors like swarming and flocking, often driven by relatively simple, reflexive responses to visual cues. In contrast, cooperative social behaviors, especially in NHPs and humans, involve greater cognitive complexity and strategic interaction. Distinguishing between imitation-based behaviors and more cognitively sophisticated, goal-directed actions is essential for refining our conceptual framework. This distinction helps clarify the underlying mechanisms of collective social behavior and sheds light on the diverse neural and cognitive processes that support different forms of group interaction. Furthermore, expanding this concept to real-life human scenarios could provide valuable insights into how the human brain processes social information and the precise neural mechanisms underlying the manifestation of psychiatric disorders when social homeostasis is disrupted (Matthews and Tye, 2019; Lee and Williams, 2025). Such advancements could have important implications for developing targeted mental health interventions. By integrating insights from diverse animal models with emerging tracking technologies, neural recording methods, and modulation techniques, we can move closer to a comprehensive understanding of collective group behavior and its dysfunction.