Retrieval-augmented generation (RAG) greatly enhances large language models (LLMs) performance in knowledge-intensive tasks. However, naive RAG methods struggle with multi-hop question answering due to their limited capacity to capture complex dependencies across documents. Recent studies employ graph-based RAG to capture document connections. However, these approaches often result in a loss of semantic coherence and introduce irrelevant noise during node matching and subgraph construction. To address these limitations, we propose NeuroPath, an LLM-driven semantic path tracking RAG framework inspired by the path navigational planning of place cells in neurobiology. It consists of two steps: Dynamic Path Tracking and Post-retrieval Completion. Dynamic Path Tracking performs goal-directed semantic path tracking and pruning over the constructed knowledge graph (KG), improving noise reduction and semantic coherence. Post-retrieval Completion further reinforces these benefits by conducting second-stage retrieval using intermediate reasoning and the original query to refine the query goal and complete missing information in the reasoning path. NeuroPath surpasses current state-of-the-art
Recent advances in artificial neural networks for machine learning, and language modeling in particular, have established a family of recurrent neural network (RNN) architectures that, unlike conventional RNNs with vector-form hidden states, use two-dimensional (2D) matrix-form hidden states. Such 2D-state RNNs, known as Fast Weight Programmers (FWPs), can be interpreted as a neural network whose synaptic weights (called fast weights) dynamically change over time as a function of input observations, and serve as short-term memory storage; corresponding synaptic weight modifications are controlled or programmed by another network (the programmer) whose parameters are trained (e.g., by gradient descent). In this Primer, we review the technical foundations of FWPs, their computational characteristics, and their connections to transformers and state space models. We also discuss connections between FWPs and models of synaptic plasticity in the brain, suggesting a convergence of natural and artificial intelligence.
The great behavioral heterogeneity observed between individuals with the same psychiatric disorder and even within one individual over time complicates both clinical practice and biomedical research. However, modern technologies are an exciting opportunity to improve behavioral characterization. Existing psychiatry methods that are qualitative or unscalable, such as patient surveys or clinical interviews, can now be collected at a greater capacity and analyzed to produce new quantitative measures. Furthermore, recent capabilities for continuous collection of passive sensor streams, such as phone GPS or smartwatch accelerometer, open avenues of novel questioning that were previously entirely unrealistic. Their temporally dense nature enables a cohesive study of real-time neural and behavioral signals. To develop comprehensive neurobiological models of psychiatric disease, it will be critical to first develop strong methods for behavioral quantification. There is huge potential in what can theoretically be captured by current technologies, but this in itself presents a large computational challenge -- one that will necessitate new data processing tools, new machine learning technique
Rapid anthropogenic environmental changes, including those due to habitat contamination, degradation, and climate change, have far-reaching effects on biological systems that may outpace animals' adaptive responses (Radchuk et al., 2019). Neurobiological systems mediate interactions between animals and their environments and evolved over millions of years to detect and respond to change. To gain an understanding of the adaptive capacity of nervous systems given and unprecedented pace of environmental change, mechanisms of physiology and behavior at the cellular and biophysical level must be examined. While behavioral changes resulting from anthropogenic activity are becoming increasingly described, identification and examination of the cellular, molecular, and circuit-level processes underlying those changes are profoundly under-explored. Hence, the field of neuroscience lacks predictive frameworks to describe which neurobiology systems may be resilient or vulnerable to rapidly changing ecosystems, or what modes of adaptation are represented in our natural world. In this review, we highlight examples of animal behavior modification and corresponding nervous system adaptation in res
This paper summarizes a research program to express the organization of sensorimotor control by specifying physiological states and the conditions for transitions among them. By a slight change in standard notation, conditional dynamics provides a moving spotlight, focussing on salient subspaces within a high-dimensional space. This mathematical approach serves as a window on the organization of sensorimotor neurobiology. The intertwined efforts to express the intrinsic organization of neurobiology and to clarify it mathematically are yielding a mathematical structure that is growing on fertile empirical ground.
The Neurobiology Of Thinking, Identity, And Geniality Abstract: Mathematically the axioms of representation are subtle, and critical. The CNS expresses its function via its internal neuronal networks in multidimensional, intrinsic frames. CNS operation can be represented by tensor transformations within and among general frames of reference. So-called metaorganization principle is transforming the dual representations of intention from one to another. The property of intentionality refers to the fact that consciousness is about objects or events, and is modulated by attention. To construct a new neural framework we must outgoing basicly from the competitive interactions among widely distributed groups of neurons. A major system in this regard is the thalamocortical system. The integrative dynamics of conscious experience suggests that the thalamocortical system behaves as a functional cluster that interacts mainly with itself.
Investigation of the neural basis of self-generated thought is moving beyond a simple identification with default network activation toward a more comprehensive view recognizing the role of the frontoparietal control network and other areas. A major task ahead is to unravel the functional roles and temporal dynamics of the widely distributed brain regions recruited during self-generated thought. We argue that various other neuroscientific methods - including lesion studies, human intracranial electrophysiology, and manipulation of neurochemistry - have much to contribute to this project. These diverse data have yet to be synthesized with the growing understanding of self-generated thought gained from neuroimaging, however. Here, we highlight several areas of ongoing inquiry and illustrate how evidence from other methodologies corroborates, complements, and clarifies findings from functional neuroimaging. Each methodology has particular strengths: functional neuroimaging reveals much about the variety of brain areas and networks reliably recruited. Lesion studies point to regions critical to generating and consciously experiencing self-generated thought. Human intracranial electroph
Dense Associative Memories or modern Hopfield networks permit storage and reliable retrieval of an exponentially large (in the dimension of feature space) number of memories. At the same time, their naive implementation is non-biological, since it seemingly requires the existence of many-body synaptic junctions between the neurons. We show that these models are effective descriptions of a more microscopic (written in terms of biological degrees of freedom) theory that has additional (hidden) neurons and only requires two-body interactions between them. For this reason our proposed microscopic theory is a valid model of large associative memory with a degree of biological plausibility. The dynamics of our network and its reduced dimensional equivalent both minimize energy (Lyapunov) functions. When certain dynamical variables (hidden neurons) are integrated out from our microscopic theory, one can recover many of the models that were previously discussed in the literature, e.g. the model presented in "Hopfield Networks is All You Need" paper. We also provide an alternative derivation of the energy function and the update rule proposed in the aforementioned paper and clarify the rela
For the Research Topic Data Assimilation and Control: Theory and Applications in Life Sciences we first review the formulation of statistical data assimilation (SDA) and discuss algorithms for exploring variational approximations to the conditional expected values of biophysical aspects of functional neural circuits. Then we report on the application of SDA to (1) the exploration of properties of individual neurons in the HVC nucleus of the avian song system, and (2) characterizing individual neurons formulated as very large scale integration (VLSI) analog circuits with a goal of building functional, biophysically realistic, VLSI representations of functional nervous systems. Networks of neurons pose a substantially greater challenge, and we comment on formulating experiments to probe the properties, especially the functional connectivity, in song command circuits within HVC.
Smoking remains the leading cause of preventable mortality worldwide. Adolescents are particularly vulnerable to the development of tobacco addiction due to ongoing brain maturation and susceptibility to social influences, such as exposure to environmental tobacco smoke (ETS). Craving -the strong desire to use drugs -already emerges with non-daily tobacco use and predicts continued use and relapse. However, the roles of craving and ETS exposure during the early stages of tobacco use in adolescence remain poorly understood. In this pre-registered study, we harness a recently developed fMRI marker of craving -the Neurobiological Craving Signature (NCS) -to compare craving-related brain responses to smoking versus social cues in adolescent Experimental Smokers (N=100) and Non-smokers (N=48) with varying levels of ETS exposure levels. Results showed that NCS responses to smoking cues compared to alternative social rewards were higher in Experimental Smokers compared to Non-smokers and predicted individual differences in self-reported craving and monthly smoking. Both smoking behavior and NCS responses were correlated with the relative amount of ETS exposure from peers compared to expos
Studying psychiatric illness has often been limited by difficulties in connecting symptoms and behavior to neurobiology. Computational psychiatry approaches promise to bridge this gap by providing formal accounts of the latent information processing changes that underlie the development and maintenance of psychiatric phenomena. Models based on these theories generate individual-level parameter estimates which can then be tested for relationships to neurobiology. In this review, we explore computational modelling approaches to one key aspect of health and illness: affect. We discuss strengths and limitations of key approaches to modelling affect, with a focus on reinforcement learning, active inference, the hierarchical gaussian filter, and drift-diffusion models. We find that, in this literature, affect is an important source of modulation in decision making, and has a bidirectional influence on how individuals infer both internal and external states. Highlighting the potential role of affect in information processing changes underlying symptom development, we extend an existing model of psychosis, where affective changes are influenced by increasing cortical noise and consequent i
Understanding the development of adolescent behavioral and mental health outcomes requires integrating genetic predisposition, environmental exposures, and neurobiological processes over time. Here, we present a unified quantitative framework that models the human body as a dynamic system, where genetic factors form the foundational state, environmental exposures act as time-varying inputs, the brain might serve as a mediation processor, and behavioral phenotypes emerge as system outputs. Using longitudinal data from the Adolescent Brain Cognitive Development (ABCD) Study, we construct harmonized multi-domain representations across six phenotypes: externalizing behavior, internalizing behavior, and four substance use initiation outcomes (alcohol, nicotine, cannabis, and any substance use). We integrate polygenic risk scores (PRS), multi-domain environmental features, and multimodal neuroimaging representations derived through stability selection and dimensionality reduction. Our framework supports both continuous longitudinal modeling and survival-based event modeling through a unified data structure. We further develop interpretable domain-level representations using principal com
While deep neural networks (DNNs) have achieved remarkable performance in tasks such as image recognition, they often struggle with generalization, learning from few examples, and continuous adaptation - abilities inherent in biological neural systems. These challenges arise due to DNNs' failure to emulate the efficient, adaptive learning mechanisms of biological networks. To address these issues, we explore the integration of neurobiologically inspired assumptions in neural network learning. This study introduces a biologically inspired learning rule that naturally integrates neurobiological principles, including sparsity, lognormal weight distributions, and adherence to Dale's law, without requiring explicit enforcement. By aligning with these core neurobiological principles, our model enhances robustness against adversarial attacks and demonstrates superior generalization, particularly in few-shot learning scenarios. Notably, integrating these constraints leads to the emergence of biologically plausible neural representations, underscoring the efficacy of incorporating neurobiological assumptions into neural network design. Preliminary results suggest that this approach could ex
Lexical Semantic Change (LSC) is the phenomenon in which the meaning of a word change over time. Most studies on LSC focus on improving the performance of estimating the degree of LSC, however, it is often difficult to interpret how the meaning of a word change. Enhancing the interpretability of LSC is a significant challenge as it could lead to novel insights in this field. To tackle this challenge, we propose a method to map the semantic space of contextualized embeddings of words obtained by a pre-trained language model to a neurobiological feature space. In the neurobiological feature space, each dimension corresponds to a primitive feature of words, and its value represents the intensity of that feature. This enables humans to interpret LSC systematically. When employed for the estimation of the degree of LSC, our method demonstrates superior performance in comparison to the majority of the previous methods. In addition, given the high interpretability of the proposed method, several analyses on LSC are carried out. The results demonstrate that our method not only discovers interesting types of LSC that have been overlooked in previous studies but also effectively searches for
Binge eating disorder (BED) is the most prevalent eating disorder. However, current diagnostic frameworks remain largely grounded in symptom-based criteria rather than underlying biological mechanisms, thereby limiting early detection and the development of biologically-informed interventions. Emerging studies have begun to investigate the neurobiological signatures of BED, yet their findings are often difficult to generalize due to the reliance on hypothesis-driven parametric models, single-modality analyses, and limited data diversity. Therefore, there is a critical need for advanced data-driven frameworks capable of modeling multimodal data to uncover generalizable and biologically meaningful signatures of BED. In this study, we propose the Interpretable Modality-Aware Mixture-of-Experts (IMA-MoE), a novel architecture designed to integrate heterogeneous neuroimaging, behavioral, hormonal, and demographic measures within a unified predictive framework. By encoding each measure as a distinct token, IMA-MoE enables flexible modeling of cross-modal dependencies while preserving modality-specific characteristics. We further introduce a token-importance mechanism to enhance interpret
Humans are inherently social and seek connection with others for survival. Recent studies suggest that acute social isolation leads to craving for social interactions, but the brain mechanisms of social craving and their relationship to brain networks underlying drug and food craving remain incompletely understood. Here we harnessed an existing dataset and tested whether the Neurobiological Craving Signature (NCS)-a recently developed fMRI-based brain-signature of drug and food craving-also predicts social craving. During fMRI, participants rated their craving for images of food, social interactions, and flowers in three different sessions: after 10h of fasting from food, 10h of social isolation, or neither (baseline; order of sessions counterbalanced). The NCS significantly predicted self-reported craving for food and social cues but not flower cues. Further, NCS responses to food were higher after fasting compared to baseline, and higher for social cues after social isolation compared to baseline, demonstrating its responsiveness to both food and social deprivation. These findings resonate with recent work showing shared brainstem circuits for hunger and social isolation, and ind
The term artificial implies an inherent dichotomy from the natural or organic. However, AI, as we know it, is a product of organic ingenuity: designed, implemented, and iteratively improved by human cognition. The very principles that underpin AI systems, from neural networks to decision-making algorithms, are inspired by the organic intelligence embedded in human neurobiology and evolutionary processes. The path from organic to artificial intelligence in digital health is neither mystical nor merely a matter of parameter count, it is fundamentally about organization and adaption. Thus, the boundaries between artificial and organic are far less distinct than the nomenclature suggests.
Large language models based Multi Agent Systems (MAS) have demonstrated promising performance for enhancing the efficiency and accuracy of code generation tasks. However,most existing methods follow a conventional sequence of planning, coding, and debugging,which contradicts the growth-driven nature of human learning process. Additionally,the frequent information interaction between multiple agents inevitably involves high computational costs. In this paper,we propose Cogito,a neurobiologically inspired multi-agent framework to enhance the problem-solving capabilities in code generation tasks with lower cost. Specifically,Cogito adopts a reverse sequence: it first undergoes debugging, then coding,and finally planning. This approach mimics human learning and development,where knowledge is acquired progressively. Accordingly,a hippocampus-like memory module with different functions is designed to work with the pipeline to provide quick retrieval in similar tasks. Through this growth-based learning model,Cogito accumulates knowledge and cognitive skills at each stage,ultimately forming a Super Role an all capable agent to perform the code generation task. Extensive experiments against
At HVEI-2012, I presented a neurobiologically-based model for trichromatic color sensations in humans, mapping the neural substrate for color sensations to V1-L4: the thalamic recipient layer of the primary visual cortex. In this paper, I propose that V1-L4 itself consists of three distinct sub-layers that directly correspond to the three primary color sensations: blue, red, and green. Furthermore, I apply this model to three aspects of color vision: the three-dimensional (3D) color solid, dichromatism, and ocular agnosticism. Regarding these aspects further: (1) 3D color solid: V1-L4 is known to exhibit a gradient of cell densities from its outermost layer (i.e., its pia side) to its innermost layer (i.e., its white matter side). Taken together with the proposition that the population size of a cell assembly directly corresponds with the magnitude of a color sensation, it can be inferred that the neurobiologically-based color solid is a tilted cuboid. (2) Chromatic color blindness: Using deuteranopia as an example, at the retinal level, M-cones are lost and replaced by L-cones. However, at the cortical level, deuteranopia manifests as a fusion of the two bottom layers of V1-L4. (3
Why do neurons communicate through spikes? By definition, spikes are all-or-none neural events which occur at continuous times. In other words, spikes are on one side binary, existing or not without further details, and on the other can occur at any asynchronous time, without the need for a centralized clock. This stands in stark contrast to the analog representation of values and the discretized timing classically used in digital processing and at the base of modern-day neural networks. As neural systems almost systematically use this so-called event-based representation in the living world, a better understanding of this phenomenon remains a fundamental challenge in neurobiology in order to better interpret the profusion of recorded data. With the growing need for intelligent embedded systems, it also emerges as a new computing paradigm to enable the efficient operation of a new class of sensors and event-based computers, called neuromorphic, which could enable significant gains in computation time and energy consumption -- a major societal issue in the era of the digital economy and global warming. In this review paper, we provide evidence from biology, theory and engineering th