Nature Neuroscience

Sporadic amyotrophic lateral sclerosis (sALS) lacks longitudinal molecular measurements, making it difficult to distinguish early disease-organising changes from downstream consequences. We present a training-free framework that extracts directional structure from static single-nucleus RNA-seq by applying discrete Hodge decomposition to gene co-expression dynamics across pseudotime-ordered donor states. The framework separates irreversible co-expression cascades from circular feedback structure and regresses out the component explained by the healthy co-expression network, allowing disease-specific organisation to be examined in isolation. Perturbation benchmarks show that experimentally imposed sources are recoverable from control-normalised off-diagonal covariance-response fields, whereas marginal variance and diagonal covariance controls do not recover the source. Applied to sALS primary motor cortex (24 donors, 10 cell types), the framework identifies oligodendrocytes as the most structurally upstream cell type and upper-motor-neuron-containing layers as the most structurally downstream (Oligo cell-type{varphi} = 0.900, with glial cell types preserving the healthy co-expression network topology, whereas neuronal cell types show collapse-dominant deformation). Cytoplasmic translation is the only pathway with reproducible cross-cell-type upstream enrichment. At the gene level, the ribosome-associated quality-control factor NEMF -- which appends C-terminal alanine-threonine tags ("CATylation") to nascent chains on stalled ribosomes -- shows disease-specific loss of co-expression coherence in seven of ten cell types despite essentially unchanged mRNA expression; the disease signal is decoupling from collision-response partners (GCN2, PKR), not expression-level change. Cross-cohort validation across three BA4 motor cortex cohorts (including two external cohorts; total N=107) reproduced the oligodendrocyte-upstream / upper-motor-neuron-downstream structural architecture (Oligo-preserved / ET-sink) in all three cohorts, with NEMF co-expression coherence loss replicated in two of three cohorts. These data support a brain-side, circuit-distal structural model in which oligodendrocyte-lineage stress occupies an upstream-like preserved compartment, while upper-motor-neuron-containing excitatory populations form a downstream sink. The pattern is consistent with -- but does not directly establish -- a cascade architecture in which oligodendrocyte stress structurally precedes motor neuron TDP-43 pathology, and would produce a clinical phenotype resembling dying-back (the conventional view of ALS, in which motor neuron pathology appears to begin at distal axons and spread retrogradely toward the cell body) yet originating centrally and glially. NEMF/CATylation network disruption is identified as a candidate intermediate structural node bridging oligodendrocyte stress and motor neuron TDP-43 pathology.

Adaptive learning requires distinguishing environmental volatility from observation stochasticity, two sources of uncertainty that demand opposite adjustments to the learning rate but inflate experienced variance similarly. Disentangling them is computationally difficult with no tractable closed-form solution. Particle-filter methods are the natural tool for this kind of joint inference, but their stochastic likelihoods and non-differentiable objectives force derivative-free fitting protocols and discourage the individual-difference analyses central to cognitive modeling, where small effect sizes leave little room for additional estimator noise. We introduce the Categorical Bayes Filter (CBF), a deterministic alternative that preserves the conditional structure of recent particle-filter accounts but replaces the stochastic outer layer with a categorical distribution on a quantile grid parameterized through differentiable Beta quantile functions. The procedure performs evidence maximization with an exact, deterministic marginal likelihood that is fully differentiable in the grid parameters. In a volatility-stochasticity task with N = 643 participants, fitted CBF dispersion parameters reveal a cross-over phenotyping pattern between volatility-blind and stochasticity-blind subjects that is not recoverable from particle-filter parameters fit to the same data under a state-of-the-art protocol. The deterministic structure also yields a trial-by-trial ambiguity signal that predicts response times not used in fitting. More broadly, the approach opens individual-level analyses in cognitive modeling and computational psychiatry that stochastic methods have effectively foreclosed.

Neurons interact at synapses, but they also communicate through physical contact and proximity, including diffusion, glia-mediated interactions, and ephaptic coupling. Standard connectomes map synapses, but cannot capture the full set of cell-cell contacts that can support these pathways. Here we extract contactomes from two large mouse visual cortex volumes at nanoscale resolution and quantify every cell-cell contact, the shared surface area of each contact, and the relationship between contact and synaptic connectivity. We find that contactomes are 5 - 10x denser than synaptic graphs, revealing that neurons physically contact a much larger set of potential neighbors than they synaptically connect to. We further find that most nearby potential neighbors are already in physical contact, indicating that local structural change would add few new candidate synaptic partners. Finally, we find that astrocytes form a single large syncytium-like network that spans the tissue and directly contacts nearly all neurons, and that glial processes lie within a micron or two of almost every synapse, indicating that synapses reside within a pervasive glia-shaped microenvironment. Together, these results show that physical contact forms a distinct layer of brain architecture that extends far beyond the synaptic connectome.

Cortical networks compute with remarkably sparse spiking activity, yet the circuit mechanisms that organize these few spikes into flexible, condition-dependent temporal codes remain poorly understood. Here we combine analyses of large-scale mouse recordings with a data-driven cortical microcircuit model (CMM) of mouse V1. In recordings, V1 activity exhibits condition-dependent spike-order sequences: peak-latency order varies with task outcome, current image identity, and preceding image identity, while remaining stable under split-half and single-trial analyses. After task optimization by backpropagation through time, the CMM reproduces this sequence-level signature and sparse activity more closely than the matched randomly connected RSNN and rate-RNN controls tested here. Ablations indicate that neuronal heterogeneity and distance-dependent local connectivity each reduce rigid sequential activity, with their combination giving the closest match to measured cortical signatures. Low-dimensional trajectory visualizations and model-silencing experiments further identify high-mutual-information early neurons whose removal perturbs task trajectories and decisions. Together, these results identify a biologically grounded computational principle: neuronal diversity and local connectivity help sparse recurrent networks avoid rigid temporal pipelines and support flexible, condition-dependent spike-order computation, providing candidate design principles for SNNs that exploit flexible temporal codes.

Opioid addiction is characterized by escalating drug use, driven in part by negative reinforcement from withdrawal, but the neural processes linking withdrawal to increased drug-taking remain poorly understood. Here, we use multisite local field potential recordings and interpretable machine learning to identify large-scale brain networks engaged by repeated opioid exposure and withdrawal. After discovering that repeated fentanyl exposure induces a progressively ramping network of widespread high beta and low gamma oscillations, we then identified a distinct brain network that selectively encodes the emergence and severity of opioid withdrawal. This network, termed EN-Withdrawal, is characterized by regional gamma oscillations and widely synchronized delta/theta oscillations. Its activity patterns predict the emergence of spontaneous and naloxone-precipitated withdrawal across multiple independent cohorts, generalizing across mice, sex, opioids, and dosing regimens, while persisting over multiple days of withdrawal. Using a novel, data-driven severity index, we find that network activity scales with individual behavioral severity without simply reflecting ongoing somatic behaviors or general aversion, suggesting that EN-Withdrawal underlies a withdrawal-induced internal state. Strikingly, network activity predicts the escalation of fentanyl self-administration on a mouse-by-mouse basis in experienced, but not drug-naive, animals. These findings reveal a neurophysiological substrate of the negative reinforcement cycle of addiction that shapes individual vulnerability.

Learning interpretable low-dimensional representations of whole-brain neuronal dynamics remains a major computational challenge in systems neuroscience. We present a wiring-agnostic deep-learning framework that couples a convolutional encoder with a temporal transformer to learn compact representations directly from volumetric calcium imaging of the entire Drosophila melanogaster brain. Trained to classify 16 experimental conditions that factorially combine metabolic state (fed, starved), sensory modality (olfaction, gustation, or combined), and stimulus valence (appetitive, aversive, or conflicting), the model organizes pan-neuronal whole-brain population activity into geometrically distinct, condition-specific clusters. Analysis of the models latent space reveals that state, modality, and valence are encoded along three near-orthogonal axes: a separable structure that emerges from the classification objective without explicit disentanglement constraints. Spatial attribution and regional importance analyses link modality decoding to distinct anatomical circuits, whereas metabolic state and valence related information show weaker regional specificity and broader distribution across the brain. Our approach does not require anatomical annotation, neuronal identification, or connectivity information, and thus provides a scalable foundation for comparative whole-brain imaging and representation learning of brain wide dynamics.

Psychedelics produce long-lasting effects, but their circuit mechanisms remain unclear. Here we show that, in rats, a single dose of lysergic acid diethylamide (LSD) persistently reduces pain affect. This effect is recapitulated by local administration in the anterior cingulate cortex (ACC), but not primary somatosensory cortex. Neuropixels recordings reveal that LSD suppresses stimulus-evoked nociceptive responses in the ACC, reducing the encoding of aversive value. Despite increasing intrinsic excitability ex vivo, LSD reduces the maximum stimulus-evoked firing of ACC neurons in vivo, indicating a dissociation between excitability and sensory encoding. Together, these findings show that psychedelics disrupt the cortical transformation of nociceptive input into aversive representations.

The human pons relays information between the brain and the body. It is affected by pathological processes, including diffuse midline gliomas (DMGs) and multiple sclerosis (MS) which predominantly arise in childhood and middle age, respectively. Although multiple studies address these disease states, a comprehensive resource for normal pons development is lacking. Here we present the first installment of PEDIA-BRAIN, an encyclopedia of gene expression and chromatin accessibility from 140,771 human pons nuclei spanning the first trimester to early adulthood, as a resource for the scientific community. Exploration of the encyclopedia identified two trajectories to mature oligodendrocytes and developmental restriction of genes for neuron to oligodendrocyte progenitor cell synapses. To illustrate the utility of the resource, we compared single cell transcriptomes from DMG and MS tissues to the encyclopedia and identified perturbation of oligodendrocyte subtypes in both diseases. Data may be accessed at https://pediabrain.nchgenomics.org.

Making decisions in complex, real-world environments is challenging. Biologically plausible strategies like reinforcement learning (RL) require attention toward reward-predictive stimuli to define task states, yet how attention and decision processes coordinate in the human brain remains unclear. We hypothesized this arises through interactions between orbitofrontal (OFC) value-based mechanisms and lateral prefrontal (LPFC) attention filtering. To test this, we combined behavioral modeling with local field potential (LFP) and single-unit recordings in 22 subjects performing a multidimensional RL task. Reward expectations were encoded in OFC and LPFC, as reflected in high-frequency LFP and OFC single-unit spiking, but modulated by attention only in LPFC. Theta LFPs encoded reward expectations and indexed attention-dependent LPFC-OFC coordination, with value-related coupling emerging pre-choice in high-attention subjects and post-choice in low-attention subjects. These findings show that prefrontal circuits dynamically coordinate to encode attention-weighted value signals, shaping state representations and providing a tractable solution to learning in complex environments.

Human language comprehension requires tracking words across intervening material to construct grammatical structures, yet two fundamental questions remain unresolved: whether maintenance of earlier words relies on sustained neural activity or activity-silent working memory mechanisms, and whether memory retrieval during integration engages domain-general or language-selective networks. Using magnetoencephalography with time-resolved decoding, we tracked main-clause subjects as participants heard sentences with embedded clauses (e.g., "The dog, who chases the cat, jumps over the mud."). The subject representation ("dog") decayed to baseline during the embedded clause but reactivated after the main-clause verb ("jumps"), with transformed rather than reinstated neural codes. Critically, reactivation emerged first in right dorsolateral prefrontal cortex before engaging frontotemporal language regions, and reactivation strength was modulated by syntactic structure and predicted comprehension accuracy. These findings demonstrate activity-silent working memory maintenance, structure-dependent retrieval, and cooperative function of domain-general and language-selective networks during hierarchical language processing.

Emergence from general anesthesia, defined by a recovery of consciousness to the wakeful state, is a clinically consequential state transition that remains a passive process dependent on drug clearance. Despite the critical use of anesthesia, the neural circuitry underlying behavioral recovery remains poorly defined. Here, we map whole-brain neural activity during emergence from isoflurane anesthesia in mice using Fos immunolabeling, tissue clearing, and light-sheet microscopy. This approach enables unbiased quantification of neural activity at cellular resolution across the intact whole brain and supports subsequent network analysis. Rather than resembling wakefulness, emergence exhibits widespread cortical suppression alongside selective activation of discrete subcortical nuclei. This pattern of activity includes both previously implicated arousal-related regions and lesser-studied structures linked to respiratory, autonomic, interoceptive, and cerebellar function. By comparing emergence to two behaviorally distinct wakeful control states, we find that control state selection substantially shapes interpretation of whole-brain activity maps. This establishes dual-state comparisons as a broadly useful strategy for state-dependent circuit mapping. Functional network analysis further elucidates candidate central regions that strongly covary together during emergence, with the most integrated region being the ventral orbital cortex. This approach allows for targeted causal investigation, linking brain-wide circuit discovery with future hypothesis-driven mechanistic interrogation. Together, we find that emergence from isoflurane anesthesia reflects selective subcortical recruitment rather than broad global reactivation toward wakefulness. Significance StatementMillions of people undergo general anesthesia each year. While anesthetic unconsciousness is induced rapidly, emergence from altered consciousness is unpredictable. Neural mechanisms that underlie behavioral emergence remain poorly defined. Using whole-brain Fos mapping at cellular resolution, we found that emergence from isoflurane anesthesia is characterized by widespread cortical suppression alongside selective activation of discrete subcortical, autonomic, hindbrain, and cerebellar nuclei. This selective systems-level activity pattern identifies behavioral emergence as more than a simple global return toward wakefulness and highlights underappreciated neural circuitry involved in post-anesthetic recovery. Network analysis of the Fos maps further identifies candidate regions for targeted causal investigation of emergence-related regions.

Midbrain circuits are thought to govern reward-guided behavior and momentary happiness, yet causal evidence in humans remains scarce. Here, we show that focal microstimulation of the substantia nigra pars compacta alters both decision making and momentary happiness in patients undergoing awake deep-brain stimulation surgery for Parkinsons disease. During a gambling task with interleaved momentary happiness ratings, stimulation delivered at outcome reduced risk-seeking in the loss domain without increasing overall gambling probability. Stimulation improved choice efficiency, guiding behavior toward gambles with higher expected value and thereby increasing cumulative earnings. Computational modeling revealed selective changes in loss processing, with reduced loss weighting and more linear loss utility, while gain processing remained unchanged. In parallel, stimulation did not tonically elevate happiness but selectively amplified outcome-related happiness changes, scaling with decision value and reward prediction errors. These results provide causal evidence that localized perturbation of human midbrain circuits reshapes valuation, choice, and momentary well-being.

New implantable-electrode fabrication strategies make dense, ultrafine electrode arrays with lower tissue burden increasingly feasible, shifting a key bottleneck for scalable brain-computer interfaces from electrode placement to readout capacity. Electrode pooling, in which multiple electrodes share a readout channel, could relax this bottleneck by combining extracellular signals before acquisition, but it has remained unclear whether such compression preserves the neural population structure needed for behavioral decoding. Here we evaluate this question using software-emulated electrode pooling in mouse sensorimotor cortex during a cue-guided reach-and-grasp task using a high-density microwire array coupled to a CMOS microelectrode-array platform. Pooled recordings retain forelimb kinematic information more effectively than a channel-matched control that discards electrodes. Pooling reduces the separability of electrode-specific spikes and sorted units, indicating a loss of some neuronal detail, but the mixed signals still preserve task-aligned low-dimensional latent dynamics that support decoding. When readout capacity is fixed, this trade-off allows broader electrode coverage to contribute to behaviorally informative population sampling. Together, these results define electrode pooling as a design trade-off for scalable readout, in which some electrode-specific neuronal information is lost but the population dynamics needed for movement decoding remain accessible.

Sensorimotor associations are typically thought to require days of training to consolidate in sensory cortex, yet adaptive behavior can emerge within minutes. Here, we developed a barrel cortex-dependent whisker-based detection task in which mice learned to associate a novel tactile whisker stimulus with reward within a single behavioral session. Longitudinal two-photon calcium imaging of layer 2/3 barrel cortex neurons revealed that reward-driven learning rapidly reorganized the neuronal representation of the whisker deflection within a single session. Population decoding tracked this transition trial-by-trial during learning with neuronal trajectories mirroring behavior. Critically, neurons that gained stimulus responsiveness across training preferentially took part in spontaneous reactivation events during learning, suggesting that online reactivations could act as a potential upstream selection mechanism. Our results suggest that reward-based learning evokes rapid sensory cortical reorganization on the timescale of minutes, which could be mediated by a concurrent reactivation-based mechanism driving plasticity.

The development of the functional cortical hierarchy, spanning sensorimotor to association systems, is exclusively studied as a function of age. During adolescence, this overlooks puberty as a major neurodevelopmental driver and source of variability. We studied sensorimotor-association axis refinement longitudinally (6323 observations across 4919 subjects), leveraging individual differences to disentangle chronological age from pubertal effects. We derived low dimensional features of sensorimotor-association axis development from resting-state functional connectomes, revealing substantial inter-individual heterogeneity in maturational trajectories that challenge group-level developmental trends and milestones. Then, we demonstrate independent effects of age and pubertal stage on sensorimotor-association axis refinement through the polarization of the cortical hierarchy. We further show that coordinated system-level shifts in network topology reflect an ongoing specialization of functional connectivity profiles across all major functional networks. Our findings frame adolescent hierarchical functional cortical maturation as an individualized, multifactorial phenomenon shaped by distinct chronological age and pubertal processes.

Animals facing threat must integrate multiple streams of information -- about danger, environment, and internal state -- into a time-pressured escape decision. In mice, this computation is performed by glutamatergic neurons of the dorsal periaqueductal grey (dPAG), but how their convergent long-range inputs combine to drive flexible decisions is unknown. Here we find that the functional weight of each input is set predominantly by the temporal statistics of its presynaptic activity, rather than by pathway identity or synaptic placement. We first used multi-region single unit recordings during naturalistic behaviour and generalised linear models to estimate the functional connectivity from midbrain, hypothalamic, and cortical inputs onto dPAG neurons. We then combined synapse-resolution circuit tracing, two-photon dendritic stimulation with whole-cell somatic and dendritic recordings, and biophysical modelling to identify the mechanisms setting these weights. We found that dPAG neurons are electrotonically compact, generating broadly uniform somatic responses to inputs across the dendritic tree. As a result, presynaptic firing dynamics -- burstiness within neurons and population synchrony -- are the dominant determinants of input efficacy. This temporal-statistics framework accounts for the measured differences in functional connectivity across input regions and predicts that input weights should change dynamically whenever presynaptic temporal structure shifts -- which we confirm by showing rapid, context-dependent reweighting of cortical input during motivational conflict. We propose that the subcellular specialisations of dPAG neurons allow them to integrate signals from distributed sources into a single decision, with input weights that can be flexibly adjusted on behavioural timescales -- a principle that may extend to other brain hubs that compute survival decisions.

The Bayesian Brain hypothesis assumes that cognition relies on internal generative models of the world, yet existing implementations remain constrained by pre-specified, task-specific generative structures and computationally heavy iterative inference schemes. Here, we introduce modular neural state-space models as a scalable realization of the Bayesian Brain, replacing fixed generative structures and pre-specified inference rules with learned world-model components and amortized neural updates. This framework preserves the core commitment to explaining observations through hidden causes while making inference learned and reusable rather than pre-specified and task-specific. Our modular implementation of these models affords learned components to be seamlessly recombined and stacked across superficially different tasks that share similar latent dynamics. Such computational reuse supports zero-shot generalization and predicts selective correlations of inference parameters between tasks. We confirm these key predictions in human behavior, identifying learned and reusable world-model components as a candidate computational principle for flexible cognition.

The ventral visual stream is widely modeled as a serial feedforward hierarchy in which V1, V4, and IT population codes develop sequentially during object recognition. We ask whether a second, concurrent coding mode exists--one organized not by anatomical order but by joint population structure across areas. Using Partial Information Decomposition applied to simultaneous multielectrode spiking recordings across all three areas at millisecond resolution--the first simultaneous three-area spiking PID analysis of the primate ventral stream--in two macaque monkeys viewing 25,000+ natural images, we decompose population coding into serial (unique per area) and synergistic (joint across areas) components at 5 ms resolution across five CNN target representations spanning low-level spatial features to high-level object identity. Three findings replicate across both animals and all five representations. First, synergistic inter-area coupling emerges before IT carries any unique object-related information--a dissociation of 15-65 ms that replicates in direction without exception across both animals--such that the joint population integrates before the apex encodes; moreover, V1-IT synergy persists for over 120 ms after V1s unique information reaches zero. Second, although V1{leftrightarrow}IT and V1{leftrightarrow}V4 coupling emerge simultaneously and rise in parallel, V1{leftrightarrow}IT exhibits stronger peak synergy at mid-to-high-level targets in both animals, suggesting a dominant role for non-serial joint coding. Third, when V1 and V4 are treated as an integrated feedforward block, their synergistic coupling with IT emerges last across all tested conditions--the feedforward foundation is the final component to join the synergistic mode, not the first. Together, these results show that serial and synergistic population codes co-occur in the same recordings, overlap in time, but follow different organizational principles, Providing a new level of nuance in our understanding of the primate ventral stream and introducing concrete constraints for biologically grounded models of vision.

The dural compartment of the meninges forms a dynamic interface between the brain and the periphery, hosting diverse immune, vascular, mural and fibroblast populations. Single-cell studies have begun charting meningeal cellular diversity, yet a comprehensive view encompassing all major cellular compartments, intercellular communication, and the influence of sex remains lacking. Here, we present a single-cell transcriptomic atlas of the adult mouse dural meninges, profiling all major cell types in male and female mice at steady state. We uncover broad sex differences in cell-type-specific proportion, transcriptional programs, intercellular communication, and disease relevant signatures. Histological and cytometry analyses validate the biological relevance of these findings, establishing this atlas as a foundation for studying meningeal contributions to neurological disease in a sex-aware manner.

Parkinsons disease is defined by dopaminergic neuron loss in the substantia nigra, yet its hallmark, exaggerated beta-band synchrony, pervades motor cortex, thalamus, and cerebellum, implicating network dynamics far beyond any single circuit. How focal subcortical dopamine depletion translates into brain-wide oscillatory pathology remains unresolved. We use a connectome-constrained multiscale model of the mouse brain, embedding biophysically detailed spiking networks of basal ganglia and cerebellum within whole-brain corticothalamic dynamics grounded in the Allen Mouse Brain Connectivity Atlas. We show that confining dopamine depletion exclusively to subcortical circuits is sufficient to produce widespread beta hypersynchrony (10-30 Hz), accompanied by heterogeneous theta and gamma dysregulation. Virtual loop ablations reveal that cortical and cerebellar beta amplification strictly requires intact cortico-basal ganglia-thalamic feedback; severing this loop confines beta to subcortical generators. These results support resonance within closed large-scale loops, rather than local rhythmogenesis, as the mechanism underlying distributed Parkinsonian beta pathology.