Karl von Frisch's ground-breaking research first demonstrated visual learning in the European honey bee (Apis mellifera). The study of Australian native bees and their cognitive abilities, however, is still a relatively new and emerging field. Here we examined visual cognition in the Australian stingless bee, Tetragonula carbonaria. First we tested for any colour preferences in T. carbonaria. Then we set stingless bees with three simple visual discrimination learning tasks using distinct colours (blue or yellow), oriented gratings (horizontal or vertical) and patterns (radial or concentric). In the colour preference task, we found evidence of a weak colour preference, with bees preferring colours blue and purple. In the visual discrimination task, T. carbonaria learned all three tasks in just 10 training trials. Bees learned equally well across the colour, orientation and pattern conditions, suggesting rapid visual learning. Future research should focus on closing knowledge gaps in Australian native bee cognition research, building upon the results of this study and exploring more complex non-elemental learning.

Transitional accounts of evolution emphasise a few changes that shape what is evolvable, with dramatic consequences for derived lineages. More recently it has been proposed that cognition might also have evolved via a series of major transitions that manipulate the structure of biological neural networks, fundamentally changing the flow of information. We used idealised models of information flow, artificial neural networks (ANNs), to evaluate whether changes in information flow in a network can yield a transitional change in cognitive performance. We compared networks with feed-forward, recurrent and laminated topologies, and tested their performance learning artificial grammars that differed in complexity, controlling for network size and resources. We documented a qualitative expansion in the types of input that recurrent networks can process compared to feed-forward networks, and a related qualitative increase in performance for learning the most complex grammars. We also noted how the difficulty in training recurrent networks poses a form of transition barrier and contingent irreversibility – other key features of evolutionary transitions. Not all changes in network topology confer a performance advantage in this task set. Laminated networks did not outperform non-laminated networks in grammar learning. Overall, our findings show how some changes in information flow can yield transitions in cognitive performance.

Honey bee colonies are facing increasing environmental stressors that threaten their health and lifespan. While the gut microbiota may play a role in honey bee physiology, the specific functions of certain bacterial species remain unclear. This study investigates whether Bombella intestini, a bacterium highly enriched in the queen gut but nearly absent in worker bees, can act as a probiotic to promote honey bee growth, metabolism, and lifespan. Our results show that B. intestini can survive in larval food and the larval gut. When larval food is inoculated with B. intestini there is increased tryptophan in both the larval diet and larval hemolymph. Bees fed this diet had a longer lifespan. This study identifies B. intestini as a potential probiotic for honey bees, providing a microbiome-based strategy to enhance their growth and longevity. These findings open new avenues for improving honey bee health management through microbial supplementation.

Bees are flexible and adaptive learners, capable of learning stimuli seen on arrival and at departure from flowers where they have fed. This gives bees the potential to learn all information associated with a feeding event, but it also presents the challenge of managing information that is irrelevant, inconsistent, or conflicting. Here, we examined how presenting bumblebees with conflicting visual information before and after feeding influenced their learning rate and what they learned. Bees were trained to feeder stations mounted in front of a computer monitor. Visual stimuli were displayed behind each feeder station on the monitor. Positively reinforced stimuli (CS +) marked feeders offering sucrose solution. Negatively reinforced stimuli (CS−) marked feeders offering quinine solution. While alighted at the feeder station the stimuli were likely not visible to the bee. The “constant stimulus” training group saw the same stimulus throughout. For the “switched stimulus” training group, the CS + changed to the CS− during feeding. Learning was slower in the “switched stimulus” training group compared to the constant stimulus” group, but the training groups did not differ in their learning performance or the extent to which they generalised their learning. The information conflict in the “switched stimulus” group did not interfere with what had been learned. Differences between the “switched” and “constant stimulus” groups were greater for bees trained on a horizontal CS + than a vertical CS + suggesting bees differ in their processing of vertically and horizontally oriented stimuli. We discuss how bumblebees might resolve this type of information conflict so effectively, drawing on the known neurobiology of their visual learning system.

Generative Pre-Trained Transformers (GPTs) are hyped to revolutionize robotics. Here we question their utility. GPTs for autonomous robotics demand enormous and costly compute, excessive training times and (often) offboard wireless control. We contrast GPT state of the art with how tiny insect brains have achieved robust autonomy with none of these constraints. We highlight lessons that can be learned from biology to enhance the utility of GPTs in robotics.

The thriving field of comparative cognition examines the behaviour of diverse animals in cognitive terms. Comparative cognition research has primarily focused on the abilities of animals — what tasks they can do — rather than on the limits of their cognition — tasks that exceed an animal’s cognitive abilities. We propose that understanding and identifying cognitive limits is as important as demonstrating the capacities of animal minds. Here, we identify challenges that have deterred the study of cognitive limits related to epistemic, practical and publication problems. The epistemic problem is concerned with how we can confidently infer a cognitive limit from null or negative results. The practical problem is how can we be certain our research has identified a cognitive limit rather than failures in tasks due to methodological or experimental design issues. The publication problem outlines the publication bias toward positive and exciting results over negative or null results in animal cognition. We propose solutions to these three challenges and examples of how to conduct research to confidently identify and confirm cognitive limits in animals. We believe a refocus on the cognitive limits of animals is the next step in the field of comparative cognition. Knowing the limits to the intelligence of different animals will aid us in appreciating the diversity of animal intelligence, and will resolve outstanding questions of how cognition evolves.

The comparative approach is a powerful way to explore the relationship between brain structure and cognitive function. Thus far, the field has been dominated by the assumption that a bigger brain somehow means better cognition. Correlations between differences in brain size or neuron number between species and differences in specific cognitive abilities exist, but these correlations are very noisy. Extreme differences exist between clades in the relationship between either brain size or neuron number and specific cognitive abilities. This means that correlations become weaker, not stronger, as the taxonomic diversity of sampled groups increases. Cognition is the outcome of neural networks. Here we propose that considering plausible neural network models will advance our understanding of the complex relationships between neuron number and different aspects of cognition. Computational modelling of networks suggests that adding pathways, or layers, or changing patterns of connectivity in a network can all have different specific consequences for cognition. Consequently, models of computational architecture can help us hypothesise how and why differences in neuron number might be related to differences in cognition. As methods in connectomics continue to improve and more structural information on animal brains becomes available, we are learning more about natural network structures in brains, and we can develop more biologically plausible models of cognitive architecture. Natural animal diversity then becomes a powerful resource to both test the assumptions of these models and explore hypotheses for how neural network structure and network size might delimit cognitive function.

Hybrid vegetable varieties have become essential for global agricultural production, offering key advantages for yield, quality and disease resistance. The production of hybrid seeds is however limited by pollination challenges, with these systems commonly associated with unattractive and low-quality floral resources, isolated growing environments and frequent insecticide applications.Here, we utilise commercial carrot seed crops to investigate the impact of hybrid pollination on the behaviour of the honey bee (Apis mellifera).Six full-strength bee colonies were equipped with solar powered radio frequency identification systems and over 900 tagged bees per season. These colonies were deployed to commercial hybrid seed crops over two consecutive seasons. The colonies were allocated to either an on-crop or off-crop group. Individual bees were autonomously monitored for the pollination period to assess key parameters such as survival, age at foraging and the number and duration of orientation and foraging trips.Hybrid carrots were found to have a significant impact on bee foraging behaviour. Bees situated on the carrot crop undertook less frequent yet longer foraging trips, resulting in less total time outside of the colony compared to off-crop bees. However, bees placed in carrot fields survived to an older age, orientated successfully, became foragers later and collected more pollen by weight, despite only 2% originating from carrot. We hope that the improved behavioural understanding can be utilised to enhance both pollinator health and hybrid seed production globally.

Laboratory studies show detrimental effects of metallic pollutants on invertebrate behaviour and cognition, even at low levels. Here we report a field study on Western honey bees exposed to metal and metalloid pollution through dusts, food and water at a historic mining site. We analysed more than 1000 bees from five apiaries along a gradient of contamination within 11 km of a former gold mine in Southern France. Bees collected close to the mine exhibited olfactory learning performances lower by 36% and heads smaller by 4%. Three-dimensional scans of bee brains showed that the olfactory centres of insects sampled close to the mine were also 4% smaller, indicating neurodevelopmental issues. Our study raises serious concerns about the health of honey bee populations in areas polluted with potentially harmful elements, particularly with arsenic, and illustrates how standard cognitive tests can be used for risk assessment.

Abstract How to identify ecological systems at risk of failure is a central question of modern biology and agriculture. Due to human impacts and global change, there is a growing need for early warning signals that identify when an ecological system is at risk of a state change before those changes become irreversible or extremely complex and costly to remediate. Honeybees (Apis mellifera) are an urgent case because our food crops heavily rely on them for pollination and annual bee colony losses reported by beekeepers across the globe are unsustainable. Given enough warning, beekeepers can rescue dying colonies, but early warning signals of death for individual bee colonies are lacking. Here we used early warning signals to investigate whether fluctuations and dynamical patterns in internal hive temperature can be used as an early indicator of impending colony failure. Across three distinct datasets we found that temperature regulation of failing colonies was different enough to distinguish them from healthy colonies weeks before they died. This signal comes early enough to intervene and assist colonies with standard beekeeping practices. Our study shows that early warning theory can help to identify practical signals of risk of state change even in systems that change state relatively rapidly, such as a dying bee colony, early enough to intervene and prevent losses.

The evolution of cognition can be understood in terms of a few major transitions-changes in the computational architecture of nervous systems that changed what cognitive capacities could be evolved by downstream lineages. We demonstrate how the idea of a major cognitive transition can be modeled in terms of where a system's effective computational architecture falls on the well-studied hierarchy of formal automata (HFA). We then use recent work connecting artificial neural networks to the HFA, which provides a way to make the structure-architecture link in natural systems. We conclude with reflections on the power and the challenges of traditional thinking when applied to neural architectures. This article is categorized under: Cognitive Biology > Evolutionary Roots of Cognition Psychology > Comparative Philosophy > Foundations of Cognitive Science.

Advancements in agricultural production have seen the rapid adoption of protected cropping systems globally. Such systems have been optimized for plant growth and efficiency, with little understanding of the potential impacts to key insect pollinators. Here, we investigate the effect of bird netting and polythene rain covers on the health and performance of honey bees (Apis mellifera L.) during the pollination of sweet cherry crops. Over two consecutive seasons, 12 full-strength colonies were equipped with tagged bees and radio frequency identification (RFID) systems. The colonies were equally divided between open control, bird netted and polythene (semi-permanent VOEN in 2019 and retractable Cravo in 2020) groups. Over 1300 individual bees were monitored for the duration of the commercial pollination period to determine behavioural parameters such as foraging commencement age, number and duration of trips and overall survival. Bees began foraging within the optimum age range (mean 15.7-24.1 days) under all covering types, with little indication of prolonged stress or increased mortality during the short season. Polythene covers (VOEN & Cravo) were found to significantly increase the total time needed for bees to orientate successfully. Once orientated, bees placed under covers conducted up to 155% more foraging trips, with a longer cumulative duration. Covering type was found to significantly impact the amount and type of pollen collected, with the most restrictive system (VOEN) yielding the highest proportion of cherry pollen. Overall, we found little evidence to suggest that the tested protective covers have a detrimental impact to honey bee foraging in cherry crops.

The evolutionary history of animal cognition appears to involve a few major transitions: major changes that opened up new phylogenetic possibilities for cognition. Here, we review and contrast current transitional accounts of cognitive evolution. We discuss how an important feature of an evolutionary transition should be that it changes what is evolvable, so that the possible phenotypic spaces before and after a transition are different. We develop an account of cognitive evolution that focuses on how selection might act on the computational architecture of nervous systems. Selection for operational efficiency or robustness can drive changes in computational architecture that then make new types of cognition evolvable. We propose five major transitions in the evolution of animal nervous systems. Each of these gave rise to a different type of computational architecture that changed the evolvability of a lineage and allowed the evolution of new cognitive capacities. Transitional accounts have value in that they allow a big-picture perspective of macroevolution by focusing on changes that have had major consequences. For cognitive evolution, however, we argue it is most useful to focus on evolutionary changes to the nervous system that changed what is evolvable, rather than to focus on specific cognitive capacities.

In a technical comment, Barron et al . ( 1 ) criticized the work of Huang et al . ( 2 ) putting the accent on the quantification of dopamine levels via high-performance liquid chromatography (HPLC), yet also including data interpretation through alternative hypotheses aimed at invalidating the original ones proposed by Huang et al . We thank the authors of this technical comment, which allows us to clarify technical aspects of our work that may have been unclear, and for promoting discussion around the conclusions of our work. Below we provide answers to the points raised in their comment.

The role of the epigenome in phenotypic plasticity is unclear presently. Here we used a multiomics approach to explore the nature of the epigenome in developing honey bee (Apis mellifera) workers and queens. Our data clearly showed distinct queen and worker epigenomic landscapes during the developmental process. Differences in gene expression between workers and queens become more extensive and more layered during the process of development. Genes known to be important for caste differentiation were more likely to be regulated by multiple epigenomic systems than other differentially expressed genes. We confirmed the importance of two candidate genes for caste differentiation by using RNAi to manipulate the expression of two genes that differed in expression between workers and queens were regulated by multiple epigenomic systems. For both genes the RNAi manipulation resulted in a decrease in weight and fewer ovarioles of newly emerged queens compared to controls. Our data show that the distinct epigenomic landscapes of worker and queen bees differentiate during the course of larval development.

The salivary gland of the black field cricket, Teleogryllus commodus Walker changed size between being starved and fed. Crickets without access to food for 72 h showed a reduction in both wet and dry mass of the glands compared with the glands from continuously fed animals at 72 h. Glands returned to size following ingestion within 10 min. Salivary glands of starved crickets (72 h) were incubated in saline containing either serotonin (5-HT) or dopamine (DA). Glands increased to pre-starvation size after 1 h incubation in situ with either 10-4 moles L-1 5-HT or 10-4 moles L-1 DA, although lower concentrations (10-5 moles L-1) did not affect gland size. From immunohistochemistry, amines appeared to shift from zymogen cells during starvation to parietal cells following feeding. High-performance liquid chromatography showed that serotonin concentration is higher than dopamine in the salivary gland removed from starved and fed crickets, but the quantity of these compounds was not dependent upon feeding state; the amine quantities increased as gland size increased. Further work is necessary to determine what might be the stimulus for gland growth and if dopamine and serotonin play a role in the stimulation of salivary gland growth after a period of starvation.

Urbanisation and industralisation has increased heavy metal pollution globally. In this chapter we consider what we currently know of how heavy metal pollutants impact insect pollinators, especially honey bees. Heavy metal pollutants are a complex ecological challenge. Heavy metals often occur as cocktails of pollutants. Both mining and the burning or refining of fossil fuels release mixtures of metal pollutants into the environment. These can cause widespread and long-lasting damage to animals at tissue, cellular and epigenomic levels. Most of what we know of the effects of heavy metal pollutants has been derived from studies of humans or mammalian model systems. What we currently know from the studies of metal pollutants of insects suggests that insects are more exposed to environmental heavy metal pollutants. There is evidence that insects bioaccumulate them quickly, and they may be harmed by levels of pollution that would be considered safe for humans. This is a serious cause for concern, but we are limited by too few studies. Taxonomic coverage of insects is poor, with studies biased toward a few pest or beneficial species, including honey bees, a few wild bee species and mosquitos. Very few studies have considered how cocktails of metal pollutants might impact insects. For these reasons we need more data to accurately determine the "safe" and harmful levels of heavy metal pollutants for insect pollinators, and establish guidelines for their protection.

Honey bee ecology demands they make both rapid and accurate assessments of which flowers are most likely to offer them nectar or pollen. To understand the mechanisms of honey bee decision-making, we examined their speed and accuracy of both flower acceptance and rejection decisions. We used a controlled flight arena that varied both the likelihood of a stimulus offering reward and punishment and the quality of evidence for stimuli. We found that the sophistication of honey bee decision-making rivalled that reported for primates. Their decisions were sensitive to both the quality and reliability of evidence. Acceptance responses had higher accuracy than rejection responses and were more sensitive to changes in available evidence and reward likelihood. Fast acceptances were more likely to be correct than slower acceptances; a phenomenon also seen in primates and indicative that the evidence threshold for a decision changes dynamically with sampling time. To investigate the minimally sufficient circuitry required for these decision-making capacities, we developed a novel model of decision-making. Our model can be mapped to known pathways in the insect brain and is neurobiologically plausible. Our model proposes a system for robust autonomous decision-making with potential application in robotics.