The cichlid Neolamprologus multifasciatus is among the smallest fish in Lake Tanganyika but can dominate areas of the lake floor, living in stable territorial groups with clear memberships that persist for years. This fish is found on 'shell bed' habitats, regions where vast numbers of empty snail shells have accumulated over millennia, preserved by the lake's alkaline waters. Group members take shelter inside the shells that they dig up from the lake floor, aggregating them within their territories with each defending its own shell fiercely, even against much larger rivals, while jointly defending the overall group territory from outsiders. Although miniscule in stature, this fish displays a rich, complex social life in which conflict and cooperation play out against the backdrop of extreme population densities, predation threats, resource defence and ingroup/outgroup interactions. The presence of these fish also imprints itself on the landscape, where their excavation efforts create a cratered, almost lunar landscape on the lake floor, the result of society level territoriality. In the ancient waters of this Lake, these fish are furious with activity, constantly tending to their shelters and engaging in social interactions comparable to the complexity and subtlety of many societies in larger taxa. (c) 2025 The Authors. Published by Elsevier Ltd on behalf of The Association for the Study of Animal Behaviour. This is an open access article under the CC BY license (http://creativecommons.org/licenses/ by/4.0/).

Abstract Spatial cognition varies widely even among closely related species, and identifying the ecological drivers of this variation remains challenging because species typically differ in multiple aspects of their ecology simultaneously. Here we performed wild experiments to compare spatial navigation in seven related species of shell-dwelling cichlids from Lake Tanganyika that share fundamental aspects of their ecology but vary in home range size and shelter availability. We displaced territorial males at successively greater distances (30 cm to 15 m) from shelters, using video tracking and automated 3D terrain reconstruction to measure return paths and homing success. We complemented this with onshore experiments examining landmark use and memory for shell locations. Of the seven species tested, three (Lamprologus ornatipinnis, L. ocellatus, Neolamprologus meeli) reliably returned to their home shells while four species (N. multifasciatus, N. brevis, N. pulcher, Telmatochromis temporalis) sought shelter in the nearest refuge. Among the navigating species, homing success declined with displacement distance, and long-distance performance (at 15 m) differed: N. meeli (largest range, fewest shelters) succeeded in 80% of trials, while L. ocellatus and L. ornatipinnis (smaller ranges, marginally more shelters) succeeded in only ~ 10% of trials. Arena experiments revealed that the two better-performing species in the field preferentially searched in areas previously associated with their home shell, with N. meeli showing additional use of experimental landmarks. Our findings provide a detailed comparative foundation for future work, highlighting how ecological context shapes spatial cognition in natural populations.

Territorial behaviour can facilitate the defence of resources but becomes disadvantageous when resources are mobile and move beyond territory boundaries. In such cases, animals may develop strategies to retain access to these resources within their territory. Here, we describe a novel behaviour in the territorial Mediterranean rainbow wrasse, Coris julis, which joins producer-scrounger aggregations with a goatfish species, Mullus surmuletus, that disturbs benthic prey during foraging. Wrasse remained within their territories and did not follow producers beyond these boundaries but repeatedly engaged in tactile interactions (brief physical contact using their ventral side and pelvic fins or caudal fin) with goatfish while they were present. Events involving tactile interactions were associated with longer feeding durations of goatfish and therefore, indirectly, a higher overall food uptake for wrasse. During these events, wrasse also had increased rates of food uptake, possibly due to the close proximity to the producer and the disturbed prey. In general, feeding with goatfish was beneficial for wrasse when compared to foraging alone. Although goatfish feeding durations were longer when tactile interactions occurred, their movement patterns through the environment were unaffected by the presence or tactile contact of wrasse. While a causal relationship between tactile contacts and length of feeding time could not be established, the underlying mechanism may involve a physiological effect on goatfish, such as by reducing stress or functioning as a reassuring safety cue. Our findings describe a novel example of tactile interactions in a marine system and may suggest a behavioural strategy that could increase access to food resources within territorial boundaries, ultimately mitigating a cost of territoriality. (c) 2026 The Author(s). Published by Elsevier Ltd on behalf of The Association for the Study of Animal Behaviour. This is an open access article under the CC BY license (http://creativecommons.org/licenses/ by/4.0/).

It has been hypothesized that divergence in animal decision processes can be attributed to differences in information processing mechanisms. However, the required comparative tests of this idea are hindered by ecological, perceptual, or motivational divergence among species that confound a direct analysis of decision processes. To overcome these limitations, we study two closely related, sympatric cichlid fishes ( Aulonocranus dewindti and Cyathopharynx furcifer ) that share visual capacities, ecological context and object-removal motivations, and we perform a systematic assessment of decisional strategy divergence in wild animals, using established paradigms from human psychology research. In a series of in situ field cognitive experiments in Lake Tanganyika, comprising over 5,000 trials, we first find that the two species make similar decisions under low perceptual load, with identical size and color preferences and comparable oddity effects. Yet, when task complexity increased and choices required integrating conflicting features, the species’ decisions diverged. C. furcifer made rapid, consistent decisions based on a single dominant feature, whereas A. dewindti took longer to decide and integrated all features of the options. A decoy paradigm confirmed that A. dewindti ’s primarily relied on available information. These findings show that closely related species can exhibit markedly different decision outcomes despite sharing sensory abilities, ecological context, and baseline preferences. The observed divergence is consistent with differences in prioritization and integration of information under complex conditions.

Across animal taxa, nest-building behavior is performed using a generalizable and flexible action sequence. In order to accomplish its goal (a stable nest), the brain appears to compare intermediate steps in the process to stored neural representations that are reminiscent of cognitive templates. Deviations from these templates drive progress, preferences, and corrections in the execution of this behavior. Here, we investigated the stereotypy vs. plasticity of nest building by the cichlid Lamprologus ocellatus, a fish species that manipulates abandoned snail shells to build shelters for breeding and protection. We find that nest building is composed of a sequence of behaviors that are tied together by a series of stimulus-response loops, allowing for restarts and shortcuts as the behavioral program unfolds. The attraction to a shell object is innate, as is the final appearance of the nest. The behavior of an inexperienced animal is initially uncoordinated but is fine-tuned by repeated building opportunities. Shells need to conform to rigid geometric criteria in order to be acceptable as a potential home. Nest building is accompanied by focused neural activity in brain regions homologous to the mammalian hippocampus and neocortex. In conclusion, we have uncovered the constraints and flexibility of an instinctive, goal-directed behavior, which appears to employ cognitive template matching. Video abstract

Background Full tissue segmentation is laborious, especially for non-model organisms, whereas accurate and reliable delineation still requires much firsthand visual inspection. A virtual environment can be equipped with suitable data representations, interaction techniques, and method interfaces as to enable the interactive delineation and quantification of anatomical structure. Situated in such an environment, analysts can benefit from reduced pre-processing, but also from in-situ learning and collaboration. Results Therefore, we apply virtual reality as a method to visualise and derive higher-level anatomical features from low-level descriptors. Following voxel-size calibration, scalable delineations and measurements are performed in virtual reality. The data representation for delineation is volume visualisation: a volume rendering or an isosurface mesh. Two delineation techniques are proposed for the placement and editing of points and segments in virtual reality. For quantification, different measures and metrics can be computed for each delineated region. To mitigate some of the fundamental challenges of virtual reality, e. g., mid-air interaction affecting precision at a distance, different virtual-reality affordances were considered as part of the design. As a result, we present Brainacle , a virtual reality application, and make its usage freely available. We incorporate Brainacle in a synchrotron tomography reconstruction pipeline to delineate and quantify the gross brain regions of 20 individuals from six species of African cichlid fish. Conclusion Brainacle , an editor for the interactive delineation and quantification of anatomical structures in virtual reality, is applicable to different biological pipelines and workflows. In particular, Brainacle can be used to quickly gain an overview of structure, ease repetitive delineation and measurement, and visually inspect and communicate findings. ### Competing Interest Statement The authors have declared no competing interest. * 2-D : two-dimensional 3-D : three-dimesional AR : augmented reality CSV file : comma-separated values file HLSL : High-Level Shader Language ID : identificator IE : immersive environment JSON file : JavaScript Object Notation file ML : machine learning MR : mixed reality MS-222 : Tricaine methanesulfonate, a chemical compound NifTI file : Neuroimaging Informatics Technology Initiative file PC : personal computer SI : Système international d’unités (International System of Units) S µ CT : synchrotron micro-computed tomography TIFF : Tagged Image File Format PBS : phosphate buffered saline, a buffer solution with pH 7.4 UI : user interface VE : virtual environment VR : virtual reality WIMP interface : windows, icons, menus, pointers interface; Deutsche Forschungsgemeinschaft, 422037984 Deutsche Forschungsgemeinschaft, 251654672 Max Planck Institute of Animal Behavior Elettra Sincrotrone Trieste, 2017009, beamline SYRMEP (SYnchrotron Radiation for MEdical Physics)

The social brain hypothesis (SBH) posits that complex social environments drive the evolution of larger brains and enlargement of specific brain regions. Among species comparisons often report contrasting relationships between social complexity and brain size, potentially due to confounding effects of phylogeny, morphology and ecology. Here, we explore this relationship in a single fish species, combining behavioural observations and brain measurements of two wild populations of the cichlid Neolamprologus brevis, which occupies similar ecological niches across its range but inhabits contrasting social environments depending on local shelter abundance. We quantified social behaviour and brain size to assess whether increased social interactions are associated with greater relative brain size or region-specific neuroanatomical adaptations. We found that individuals from the more socially complex population, which exhibited more frequent social interactions, had significantly larger total brain volumes compared to the less social population. We also found that the more social population exhibited relatively larger telencephalon and smaller hypothalamus volumes, suggesting mosaic adaptation to social demands. Feeding behaviour did not differ between populations, suggesting that differences in energy intake are unlikely to account for brain size variation. By integrating behavioural and neuroanatomical data, our study provides empirical support for the SBH in a natural, within-species comparison.

Resource heterogeneity is a widespread phenomenon, as resources are rarely spaced evenly across a landscape. Variation in resource density and distribution can have a myriad of behavioural, ecological, and evolutionary consequences for populations, yet clarifying these effects is still challenging. We combine both novel and previously published data on genetic parentage, relatedness, life history, and predation to present a comprehensive field study of a shell bed in Lake Tanganyika. Here, a wild population of the cichlid fish Neolamprologus multifasciatus is naturally subdivided into habitat regions that differ immensely in shelter density and distribution, as well as in the capacity for the fish to physically rearrange their shelters into clusters (i.e., engage in niche construction). Shelters were evenly, densely, and continuously spaced in one habitat, while they were highly clustered in the other habitat. We expected the environmental potential for polygyny to be greater in the clustered habitat relative to the continuous habitat. Predation regimes and life history traits differed, with N. multifasciatus in the evenly distributed habitat experiencing higher predation threats, earlier maturation, and slower growth than those in the clustered habitat. Metrics of selection, however, were surprisingly consistent between the two habitats, as were patterns of dispersal. Overall, our research leverages the natural subdivision of a wild population into distinct habitats to investigate the ecological and evolutionary implications of resource heterogeneity and habitat modification.

Persons use and respond to various cues and behaviors as part of their daily communication and information exchange with others. When these are obscured or obstructed by the technology, users tend to experience rifts in collaboration. For this reason, we discuss different group factors and their role in collaboration in immersive environments. By including perspectives from visual and immersive analytics, embodied interaction, social psychology, and collective behavior, we identify common motives in literature that deserve the designer's attention. We list six considerations for designing and evaluating collaboration in immersive environments. The collection of these considerations could serve to inform future research on the topic.

Many animal species have been shown to discriminate between individual humans in captive settings and may use a variety of cues to do so. Empirical evidence remains scarce for animals in the wild, however, particularly in aquatic contexts. For the first time, we investigated discrimination of individual humans by fish in the wild. We first trained two species of fish, saddled sea bream Oblada melanura and black sea bream Spondyliosoma cantharus , to follow a human diver to obtain a food reward. We then investigated whether they could discriminate between two human divers and follow the correct one in an operant-conditioning paradigm. We show that both species were able to quickly learn to discriminate between the two divers when they wore different diving gear. However, they showed no preference when both divers wore identical gear, suggesting that discrimination is based predominantly on visual cues from the dive gear. We discuss the implications of these results for ethical considerations and research practices.

Behavioral biologists, from neuroscientists to ethologists, rely on observation and scoring of behavior. In the past decade, numerous methods have emerged to automate this scoring through machine learning approaches. Yet, these methods are typically specified towards laboratory settings with only two animals, or employed in cases with well-separated behavioral categories. Here, we introduce the vassi Python package, focusing on supervised classification of directed social interactions and cases in which continuous variation in behavior means categories are less distinct. Our package is broadly applicable across species and social settings, including single individuals, pairs and groups, and implements a validation tool to separate behavioral edge cases. vassi has comparable performance to existing approaches on a behavioral classification benchmark, the CALMS21 mouse resident-intruder dataset, and we demonstrate its applicability on a novel, more naturalistic and complex dataset of cichlid fish groups. Our approach highlights future challenges in extending supervised behavioral classification to more naturalistic settings, and offers a methodological framework to overcome these challenges.

Artificial light at night (ALAN) can disrupt the natural behaviour, physiology, and circadian rhythms of organisms exposed to it, and therefore presents a significant and widespread ecological concern. ALAN typically comprises a wide range of wavelengths, and different wavelengths have different effects on circadian clocks. In the animals investigated thus far, short and middle wavelengths are intensely involved in synchronisation and entrainment, but we still have a poor understanding of how different wavelengths might affect behaviour when animals are exposed to ALAN, in particular whether some wavelengths are disproportionally detrimental. This experiment examined the direct and transgenerational effects of 10 different wavelength treatments of ALAN on behaviour in zebrafish ( Danio rerio), ), a diurnally active model organism. Across a 10-day period, female zebrafish were exposed to either a monochromatic wavelength, white light ALAN, or to a control treatment, and the individual impacts of each treatment on locomotion and anxiety-like behaviours were examined both for solitary fish and fish in groups. We found the strongest impact at short wavelengths (365 to 470 nm), with individuals and groups of zebrafish showing more anxiety-like behaviour after fewer nights of ALAN exposure relative to the other wavelengths. Furthermore, F1 offspring born from ALAN-exposed mothers displayed less frequent movement and shorter movement distances despite never being exposed to ALAN themselves, regardless of the spectral treatment. Our results highlight both the specific and broad-spectrum potential for ALAN to cause disruption to locomotion in adult zebrafish and their offspring.

A central challenge in understanding the evolution of cognition is the ability to compare a set of species differing in a trait of interest while being ecologically and phylogenetically close. Here, we examine whether differences in bower-building flexibility are related to differences in cognitive flexibility between two Tanganyikan cichlids. Cognitive flexibility enables animals to modify their decision rules when faced with new situations, and inhibitory control, the ability to inhibit a normally favoured response, is an essential component of this capacity. We tested male Aulonocranus dewindti and Cyathopharynx furcifer in a choice-against-preference paradigm. Both species clean their bowers of foreign objects and we found that both preferred to remove a snail shell over a stone. We tested their ability to modify this preference and learned to preferably select the stone instead of the shell. Although neither species showed clear learning of the new preference rule, both demonstrated inhibitory control through increased decision times and manipulations of the objects when selecting the stone. Specifically, A. dewindti, the species exhibiting greater behavioural flexibility in the construction of their bowers, selected the stone in fewer trials than C. furcifer, providing support for a link between behavioural flexibility in bower construction and cognitive flexibility.

Data in behavioral research is often quantified with event-logging software, generating large data sets containing detailed information about subjects, recipients, and the duration of behaviors. Exploring and analyzing such large data sets can be challenging without tools to visualize behavioral interactions between individuals or transitions between behavioral states, yet software that can adequately visualize complex behavioral data sets is rare. TIBA (The Interactive Behavior Analyzer) is a web application for behavioral data visualization, which provides a series of interactive visualizations, including the temporal occurrences of behavioral events, the number and direction of interactions between individuals, the behavioral transitions and their respective transitional frequencies, as well as the visual and algorithmic comparison of the latter across data sets. It can therefore be applied to visualize behavior across individuals, species, or contexts. Several filtering options (selection of behaviors and individuals) together with options to set node and edge properties (in the network drawings) allow for interactive customization of the output drawings, which can also be downloaded afterwards. TIBA accepts data outputs from popular logging software and is implemented in Python and JavaScript, with all current browsers supported. The web application and usage instructions are available at tiba.inf.uni-konstanz.de. The source code is publicly available on GitHub: github.com/LSI-UniKonstanz/tiba.

1. To maximize their fitness, plants have to adjust their allocation strategy according to their abiotic and biotic environments. Plants can use the ratio of red to far-red light (R:FR) to sense neighbours, allowing them to modify their growth in response to aboveground competition. 2. In this study, we used supplemental FR light to artificially lower the R:FR of the lower leaves of common sunflowers (Helianthus annuus) to examine how plants change their growth in response to the threat of neighbours. We combined this treatment with a nitrogen fertilization treatment to investigate how responses to neighbour-detection interact with nitrogen limitation. 3. Plants grown in low R:FR increased in height at the expense of root growth, resulting in nitrogen limitation that restricted leaf growth. However, we found that plants reduced their nitrogen investment into leaves in low R:FR. By weakening the nitrogen sink strength of these lower leaves before they experienced low photosynthetically active radiation, plants were able to preemptively allocate nitrogen to leaves higher in the canopy. 4. Plants responded to the perception of neighbours by simultaneously diverting resources from root growth to stem elongation and from leaves threatened by neighbours to leaves that would pose a threat to neighbours. This whole-plant response to neighbour-detection enables plants to change their allocation in a way that simultaneously manages their limited nitrogen and prepares them for future light competition.

The Social Brain Hypothesis (SBH) proposes that complex social environments drive the evolution of larger brains and specific neuroanatomical adaptations. This relationship can be difficult to study in the wild, because species that differ in social organization may also diverge in morphology, ecology, phylogeny, and other life history parameters. Here we use two populations of the shell-dwelling cichlid Neolamprologus brevis with contrasting social environments to test whether increased social complexity is associated with larger brain sizes or specific regional adaptations. Behavioral observations revealed similarly low feeding rates in both populations, but significantly more frequent social interaction frequencies in the population one of the populations. This population had larger total brain volumes relative to body size, with a disproportionately larger telencephalon and a smaller hypothalamus, suggesting region-specific adaptations to social demands. By integrating behavioral quantification and neuroanatomical analysis, our study highlights the importance of sociality as a driver of brain evolution and demonstrates the utility of cichlid fish as a model for testing the SBH in non-mammalian systems. These findings provide empirical support for the SBH and underscore the value of combining behavioral and morphological data in evolutionary neuroethology. ### Competing Interest Statement The authors have declared no competing interest.

Many studies have found a link between higher habitat structural complexity and increased relative brain size in vertebrates. Here we explore this relationship in a multi-species comparison, comparing ten species of wild cichlids that differ in their social and territorial behaviour, but which occur across four ecologically similar but structurally diverse rocky habitats. This design allows us to perform repeated intra-specific comparisons, avoiding confounds associated with comparisons across species boundaries. We sampled 147 fish, analysing brain size and architecture while controlling for body mass and species-specific variability and compared this with habitat complexity, quantified using underwater video and three-dimensional reconstructions. Our results challenge the Clever Foraging Hypothesis (CFH), which posits that greater habitat complexity correlates with larger brain sizes. Contrary to CFH, fish from the least complex habitat had significantly larger brains. Additionally, brain architecture analysis indicated a significant enlargement of the cerebellum in fish from less complex habitats, whereas the hypothalamus showed a non-significant negative trend. Taken together, these results indicate that lower habitat complexity may impose higher cognitive demands on spatial memory and navigation due to limited refuges and increased predation risk. This study highlights the need to reconsider the assumed linear positive relationship between environmental complexity and brain development, suggesting that simpler environments might also impose significant cognitive and ecological challenges that drive brain evolution. Our findings underscore the importance of considering intra-species variability and the specific ecological and cognitive demands of different habitats in studies of brain evolution. ### Competing Interest Statement The authors have declared no competing interest.

Abstract Resource quality is an important concept in ecology and evolution that attempts to capture the fitness benefits a resource affords to an organism. Yet “quality” is a multivariate concept, potentially affected by many variables pertaining to the resource, its surroundings, and the resource chooser. Researchers often use a small number of proxy variables to simplify their estimation of resource quality, but without vetting their proxies against a wider set of potential quality estimators this approach risks overlooking potentially important characteristics that can explain patterns of resource use in their study systems. Here we used Neolamprologus multifasciatus, a group‐living cichlid fish that utilizes empty snail shells as shelter resources, to examine how shells were used by, and partitioned among, group members in relation to a range of attributes, including shell size, intactness, texture, spatial position, and usage by heterospecifics. This approach generated a comprehensive picture of what characteristics contribute to the attractiveness and quality of each shell resource, confirming the importance of two previously proposed shell characteristics, size and intactness, but highlighting the influences of other unexplored variables, including shell spatial position and usage by heterospecifics. We also present a generally applicable “resource attractiveness index” as a means to estimate resource quality based on resource choice data. This index incorporates information from any number of resource characteristics and is of particular use when researchers wish to quantify resource value, but many characteristics jointly contribute to the value and attractiveness of the resource.