We introduce CheeseBench, a benchmark that evaluates large language models (LLMs) on nine classical behavioral neuroscience paradigms (Morris water maze, Barnes maze, T-maze, radial arm maze, star maze, operant chamber, shuttle box, conditioned place preference, and delayed non-match to sample), spanning six cognitive dimensions. Each task is grounded in peer-reviewed rodent protocols with approximate animal baselines. The agent receives a unified system prompt with no task-specific instructions and must discover goals purely from ASCII text observations and reward signals, much like a rodent placed into an unfamiliar apparatus. We evaluate six open-weight LLMs (3B to 72B parameters) on text-based ASCII renderings and compare against both a random baseline and a graph-based reinforcement learning agent. Our best model (Qwen2.5-VL-7B) reaches 52.6% average success on ASCII input, compared to 32.1% for random agents and 78.9% for approximate rodent baselines. We find that (1) scaling beyond 7B yields diminishing returns, (2) longer context history degrades performance, (3) chain-of-thought prompting hurts rather than helps, and (4) a vision-language architecture provides an advantage
Scientific discovery pipelines typically involve complex, rigid, and time-consuming processes, from data preparation to analyzing and interpreting findings. Recent advances in AI have the potential to transform such pipelines in a way that domain experts can focus on interpreting and understanding findings, rather than debugging rigid pipelines or manually annotating data. As part of an active collaboration between data science/AI researchers and behavioral neuroscientists, we showcase an example AI-enhanced pipeline, specifically designed to transform and accelerate the way that the domain experts in the team are able to gain insights out of experimental data. The application at hand is in the domain of behavioral neuroscience, studying fear generalization in mice, an important problem whose progress can advance our understanding of clinically significant and often debilitating conditions such as PTSD (Post-Traumatic Stress Disorder). We identify the emerging paradigm of "In-Context Learning" (ICL) as a suitable interface for domain experts to automate parts of their pipeline without the need for or familiarity with AI model training and fine-tuning, and showcase its remarkable ef
Understanding how decision making changes across the lifespan is a central challenge for neuroscience, yet research on cognitive aging has remained largely disconnected from the theoretical and computational advances that now shape modern systems neuroscience. Over the past two decades, theoretical frameworks have transformed how we study cognition in young, healthy brains, providing principled tools to model latent decision states, neural dynamics, population codes, and interareal communication. In contrast, aging research has often relied on single metric behavioral readouts, cross sectional comparisons, and descriptive neural analyses, limiting our ability to explain fundamental differences in individual aging trajectories. This gap represents a missed opportunity because aging offers a powerful platform for testing theories of neural computation, stability, and flexibility under changing biological constraints. Here, we argue that closer integration between aging research and contemporary theoretical neuroscience can move the field beyond descriptive accounts toward more mechanistic explanations of decision making across the lifespan. To this end, we outline how recent advances
Modern neuroscience relies on integrating multi-scale, multimodal datasets to uncover the neural principles underlying intelligence. However, analytical challenges posed by highly heterogeneous data and fragmented workflows increasingly constrain discoveries. Here we introduce SeekBrain, an autonomous multi-agent framework designed to accelerate neuroscience discovery through domain-grounded hierarchical planning and cross-modal data analysis. SeekBrain dynamically constructs a repertoire of analysis recipes extracted from code-paper pairs. By coupling this codified expertise with agentic planning and execution engines, the framework scalably generates hypotheses and analytical pipelines on demand. Systematic evaluation on the expert-annotated BrainArena benchmark demonstrates that SeekBrain substantially outperforms state-of-the-art agent baselines across various analysis tasks. Crucially, when deployed in real-world research, SeekBrain integrated behavioral, neural, and anatomical data to reveal structured, distributed neural representations of larval zebrafish behavior and a shared axis of regional decoding strength across the brain in a mouse decision-making task. These results
Behavioral evaluation is the dominant paradigm for assessing alignment in large language models (LLMs). In current practice, observed compliance under finite evaluation protocols is treated as evidence of latent alignment. However, the inference from bounded behavioral evidence to claims about global latent properties is rarely analyzed as an identifiability problem. In this paper, we study alignment evaluation through the lens of statistical identifiability under partial observability. We allow agent policies to condition their behavior on observable signals correlated with the evaluation regime, a phenomenon we term evaluation awareness. Within this framework, we formalize the Alignment Verifiability Problem and introduce Normative Indistinguishability, which arises when distinct latent alignment hypotheses induce identical distributions over evaluator-accessible observations. Our main theoretical contribution is a conditional impossibility result: under finite behavioral evaluation and evaluation-aware policies, observed compliance does not uniquely identify latent alignment, but only membership in an equivalence class of conditionally compliant policies, under explicit assumpti
Current AI alignment paradigms rely on behavioral correction: external supervisors (e.g., RLHF) observe outputs, judge against preferences, and adjust parameters. This paper argues that behavioral correction is structurally analogous to an economy without property rights, where order requires perpetual policing and does not scale. Drawing on institutional economics (Coase, Alchian, Cheung), capability mutual exclusivity, and competitive cost discovery, we propose alignment as institutional design: the designer specifies internal transaction structures (module boundaries, competition topologies, cost-feedback loops) such that aligned behavior emerges as the lowest-cost strategy for each component. We identify three irreducible levels of human intervention (structural, parametric, monitorial) and show that this framework transforms alignment from a behavioral control problem into a political-economy problem. No institution eliminates self-interest or guarantees optimality; the best design makes misalignment costly, detectable, and correctable. We conclude that the proper goal is institutional robustness-a dynamic, self-correcting process under human oversight, not perfection. This wo
Autonomous vehicles (AVs) have significantly advanced in real-world deployment in recent years, yet safety continues to be a critical barrier to widespread adoption. Traditional functional safety approaches, which primarily verify the reliability, robustness, and adequacy of AV hardware and software systems from a vehicle-centric perspective, do not sufficiently address the AV's broader interactions and behavioral impact on the surrounding traffic environment. To overcome this limitation, we propose a paradigm shift toward behavioral safety, a comprehensive approach focused on evaluating AV responses and interactions within traffic environment. To systematically assess behavioral safety, we introduce a third-party AV safety assessment framework comprising two complementary evaluation components: Driver Licensing Test and Driving Intelligence Test. The Driver Licensing Test evaluates AV's reactive behaviors under controlled scenarios, ensuring basic behavioral competency. In contrast, the Driving Intelligence Test assesses AV's interactive behaviors within naturalistic traffic conditions, quantifying the frequency of safety-critical events to deliver statistically meaningful safety
Modern AI models, such as large language models, are usually trained once on a huge corpus of data, potentially fine-tuned for a specific task, and then deployed with fixed parameters. Their training is costly, slow, and gradual, requiring billions of repetitions. In stark contrast, animals continuously adapt to the ever-changing contingencies in their environments. This is particularly important for social species, where behavioral policies and reward outcomes may frequently change in interaction with peers. The underlying computational processes are often marked by rapid shifts in an animal's behaviour and rather sudden transitions in neuronal population activity. Such computational capacities are of growing importance for AI systems operating in the real world, like those guiding robots or autonomous vehicles, or for agentic AI interacting with humans online. Can AI learn from neuroscience? This Perspective explores this question, integrating the literature on continual and in-context learning in AI with the neuroscience of learning on behavioral tasks with shifting rules, reward probabilities, or outcomes. We will outline an agenda for how specifically insights from neuroscienc
Digital traces of daily activities, such as e-commerce (EC) purchase histories, provide scalable signals for public health surveillance, yet their epidemiological validity remains unclear. This study validates a behavioral proxy for disease onset, defined as transitions from regular to therapeutic diets, by comparing large-scale EC data (N=55,645) against independent insurance-derived clinical records. Using feline lower urinary tract disease (FLUTD) as a case study, the proxy showed strong agreement with clinical data for ingredient-level risk patterns (r=0.74) and seasonal dynamics (r=0.82). Furthermore, analysis using EC data alone reproduced the established protective association of wet food consumption. These results demonstrate that validated behavioral signals from EC data can serve as cost-effective complements to traditional surveillance, with potential applicability to monitoring lifestyle-related diseases in human populations.
Connecting neural activity to function is a common aim in neuroscience. How to define and conceptualize function, however, can vary. Here I focus on grounding this goal in the specific question of how a given change in behavior is produced by a change in neural circuits or activity. Artificial neural network models offer a particularly fruitful format for tackling such questions because they use neural mechanisms to perform complex transformations and produce appropriate behavior. Therefore, they can be a means of causally testing the extent to which a neural change can be responsible for an experimentally observed behavioral change. Furthermore, because the field of interpretability in artificial intelligence has similar aims, neuroscientists can look to interpretability methods for new ways of identifying neural features that drive performance and behaviors.
The exponential growth of neuroscience literature presents a significant challenge for researchers seeking to efficiently access and utilize relevant information. To address this issue, we introduce the Brain Knowledge Engine (BrainKnow), an automated system designed to extract, link, and synthesize neuroscience knowledge from scientific publications. BrainKnow constructs a comprehensive knowledge graph encompassing 3,626,931 relationships across 37,011 neuroscience concepts, derived from 1,817,744 articles. This vast repository of knowledge is accessible through a user-friendly web interface, facilitating efficient navigation and data retrieval. BrainKnow employs advanced graph network algorithms, specifically Node2Vec, to enhance knowledge recommendation and visualization. This enables users to explore semantic relationships between concepts, predict potential new relationships, and gain a deeper understanding of the interconnectedness within neuroscience. Additionally, BrainKnow ensures real-time updates by synchronizing with PubMed, providing researchers with access to the most current information. BrainKnow serves as a valuable resource for neuroscience researchers, offering a
Endowing brain anatomy, dynamics, and function with a network structure is becoming standard in neuroscience. In its simplest form, a network is a collection of units and relationships between them. The pattern of relations among the units encodes numerous properties which have been shown to have a profound effect on networked systems' dynamics and function. In an effort to strike a balance between idealization and detail, network neuroscience studies typically involve simplifying assumptions at both neural and network modeling levels. However, the extent to which existing neural models depend on such approximations is as yet poorly understood. Here, we discuss whether and how increasing neurophysiological detail and generalizing the basic simple network structure often adopted in network neuroscience may help improve our understanding of brain phenomenology and function.
In cognitive neuroscience research, Representational Dissimilarity Matrices (RDMs) are often incomplete because pairwise similarity judgments cannot always be exhaustively collected as the number of pairs rapidly increases with the number of conditions. Existing methods to fill these missing values, such as deep neural network imputation, are powerful but computationally demanding and relatively opaque. We introduce a simple algorithm based on geometric inference that fills missing dissimilarity matrix entries using known distances. We use tests on publicly available empirical cognitive neuroscience datasets, as well as simulations, to demonstrate the method's effectiveness and robustness across varying sparsity and matrix sizes. We have made this geometric reconstruction algorithm, implemented in Python and MATLAB, publicly available. This method provides a fast and accurate solution for completing partial dissimilarity matrices in the cognitive neurosciences.
The deployment of large language models (LLMs) in diverse applications requires a thorough understanding of their decision-making strategies and behavioral patterns. As a supplement to a recent study on the behavioral Turing test, this paper presents a comprehensive analysis of five leading LLM-based chatbot families as they navigate a series of behavioral economics games. By benchmarking these AI chatbots, we aim to uncover and document both common and distinct behavioral patterns across a range of scenarios. The findings provide valuable insights into the strategic preferences of each LLM, highlighting potential implications for their deployment in critical decision-making roles.
We present the case for developing the tardigrade (Hypsibius exemplaris) into a model organism for systems neuroscience. These microscopic, transparent animals (~300-500 microns) are among the smallest known to possess both limbs (eight) and eyes (two), with a nervous system of only a few hundred neurons organized into a multi-lobed brain, ventral nerve cord, and a series of ganglia along the body. Despite their neuroanatomical simplicity, tardigrades exhibit complex behaviors, including multi-limbed walking gaits, individual limb grasping, phototaxis, and transitions between active and dormant states. These behaviors position tardigrades as a uniquely powerful system for addressing certain fundamental questions in systems neuroscience, such as: How do nervous systems coordinate multi-limbed behaviors? How are top-down and bottom-up motor control systems integrated? How is stereovision-guided navigation implemented? What mechanisms underlie neural resilience and recovery during environmental stress? We review current knowledge of tardigrade neuroanatomy, behavior, and genomics, and we identify opportunities and challenges for leveraging their unique biology. We propose developing e
The proliferation and refinement of affordable virtual reality (VR) technologies and wearable sensors have opened new frontiers in cognitive and behavioral neuroscience. This chapter offers a broad overview of VR for anyone interested in leveraging it as a research tool. In the first section, it examines the fundamental functionalities of VR and outlines important considerations that inform the development of immersive content that stimulates the senses. In the second section, the focus of the discussion shifts to the implementation of VR in the context of the neuroscience lab. Practical advice is offered on adapting commercial, off-theshelf devices to specific research purposes. Further, methods are explored for recording, synchronizing, and fusing heterogeneous forms of data obtained through the VR system or add-on sensors, as well as for labeling events and capturing game play.
This paper provides a perspective on applying the concepts of information thermodynamics, developed recently in non-equilibrium statistical physics, to problems in theoretical neuroscience. Historically, information and energy in neuroscience have been treated separately, in contrast to physics approaches, where the relationship of entropy production with heat is a central idea. It is argued here that also in neural systems information and energy can be considered within the same theoretical framework. Starting from basic ideas of thermodynamics and information theory on a classic Brownian particle, it is shown how noisy neural networks can infer its probabilistic motion. The decoding of the particle motion by neurons is performed with some accuracy and it has some energy cost, and both can be determined using information thermodynamics. In a similar fashion, we also discuss how neural networks in the brain can learn the particle velocity, and maintain that information in the weights of plastic synapses from a physical point of view. Generally, it is shown how the framework of stochastic and information thermodynamics can be used practically to study neural inference, learning, and
Neuroscience research has evolved to generate increasingly large and complex experimental data sets, and advanced data science tools are taking on central roles in neuroscience research. Neurodata Without Borders (NWB), a standard language for neurophysiology data, has recently emerged as a powerful solution for data management, analysis, and sharing. We here discuss our efforts to implement NWB data science pipelines. We describe general principles and specific use cases that illustrate successes, challenges, and non-trivial decisions in software engineering. We hope that our experience can provide guidance for the neuroscience community and help bridge the gap between experimental neuroscience and data science.
Pharmacies are critical in healthcare systems, particularly in low- and middle-income countries. Procuring pharmacists with the right behavioral interventions or nudges can enhance their skills, public health awareness, and pharmacy inventory management, ensuring access to essential medicines that ultimately benefit their patients. We introduce a reinforcement learning operational system to deliver personalized behavioral interventions through mobile health applications. We illustrate its potential by discussing a series of initial experiments run with SwipeRx, an all-in-one app for pharmacists, including B2B e-commerce, in Indonesia. The proposed method has broader applications extending beyond pharmacy operations to optimize healthcare delivery.
A new scientific field is introduced and discussed, named cybernetical neuroscience, which studies mathematical models adopted in computational neuroscience by methods of cybernetics -- the science of control and communication in a living organism, machine and society. It also considers the practical application of the results obtained when studying mathematical models. The main tasks and methods, as well as some results of cybernetic neuroscience are considered.