Nature Neuroscience

An important current question in neuroimaging concerns the sample sizes required for producing reliable and reproducible results. Recent findings suggest that brain-wide association studies (BWAS) linking neuroimaging features with behavioural phenotypes in the general population are characterised by (very) weak effects and consequently need large samples sizes of 3000+ to lead to reproducible findings. A second, important goal in neuroimaging is to study brain structure and function under disease conditions, where effects are likely much larger. This difference in effect size is important. We show by means of power calculations and empirical analysis that neuroimaging studies in clinical populations need hundreds -and not necessarily thousands-of participants to lead to reproducible findings.

Alcohol use disorder (AUD) induces complex transcriptional and regulatory changes across multiple brain regions including the caudate nucleus, which remains understudied. Using paired single-nucleus RNA-seq and ATAC-seq on caudate samples from 143 human postmortem brains, including 74 with AUD, we identified 17 distinct cell types. We found that a significant portion of the alcohol-induced changes in gene expression occurred through altered chromatin accessibility. Notably, we identified novel transcriptional and chromatin accessibility differences in medium spiny neurons, impacting pathways such as RNA metabolism and immune response. A small cluster of D1/D2 hybrid neurons showed distinct differences, suggesting a unique role in AUD. Microglia exhibited distinct activation states deviating from classical M1/M2 designations, and astrocytes entered a reactive state partially regulated by JUND, affecting glutamatergic synapse pathways. Oligodendrocyte dysregulation, driven in part by OLIG2, was linked to demyelination and increased TGF-{beta}1 signaling from microglia and astrocytes. We also observed increased microglia-astrocyte communication via the IL-1{beta} pathway. Leveraging our multiomic data, we performed cell type-specific expression quantitative trait loci analysis, integrating that with public genome-wide association studies to identify AUD risk genes such as ADAL and PPP2R3C, providing a direct link between genetic variants, chromatin accessibility, and gene expression in AUD. These findings not only provide new insights into the genetic and cellular mechanisms in the caudate related to AUD but also demonstrate the broader utility of large-scale multiomic studies in uncovering complex gene regulation across diverse cell types, which has implications beyond the substance use field.

SummaryMyelin loss in the central nervous system can cause permanent motor or cognitive deficits in patients with multiple sclerosis (MS). While current immunotherapy treatments decrease the frequency of demyelinating episodes, they do not promote myelin repair or functional recovery. Vagus nerve stimulation (VNS) is a neuromodulation therapy which enhances neuroplasticity and the recovery of motor function after stroke, but its effects on myelin repair are not known. To determine if VNS influences myelin repair, we applied VNS following a demyelinating injury and measured longitudinal myelin dynamics and functional recovery. We found that VNS promotes remyelination by increasing the generation of myelinating oligodendrocytes. Pairing VNS with a skilled reach task leads to the regeneration of myelin sheaths on previously myelinated axon segments, enhancing the restoration of the original pattern of myelination. Moreover, the magnitude of sheath pattern restoration correlates with long-term motor functional improvement. Together, these results suggest that recovery of the myelin sheath pattern is a key factor in the restoration of motor function following myelin loss and identify paired VNS as a potential remyelination therapy to treat demyelinating diseases.

We recently described astrocyte pro-inflammatory epigenetic memory based on multiple complementary in vivo and in vitro studies, and the analysis of multiple sclerosis samples. Based on bioinformatic analyses, ODea and Liddelow argued that the astrocyte epigenetic memory we described is the result of contamination with immune cells, particularly myeloid cells. We rebut ODea and Liddelow arguments as follows: (1) We show substantial purity of astrocytes analyzed in in vivo and in vitro systems; (2) We recapitulate astrocyte memory responses using five independent pure astrocyte in vitro systems, and show its dependency on the histone acetyl transferase p300; and (3) Using the Liddelow lab bioinformatic pipeline to implement purity and cell-quality criteria, we detect astrocyte epigenetic memory in five independent scRNA-seq experimental autoimmune encephalomyelitis (EAE) and multiple sclerosis (MS) astrocyte datasets. These additional analyses and studies provide further support for the existence of astrocyte pro-inflammatory epigenetic memory.

The ability to retrieve a single episode encountered just once is a hallmark of human intelligence and episodic memory[1]. Yet, decoding a specific memory from neuronal activity in the human brain remains a formidable challenge. Here, we develop a transformer neural network model[2, 3] trained on neuronal spikes from intracranial microelectrodes recorded during a single viewing of an audiovisual episode. Combining spikes throughout the brain via cross-channel attention[4], capable of discovering neural patterns spread across brain regions and timescales, individual participant models predict memory retrieval of specific concepts such as persons or places. Brain regions differentially contribute to memory decoding before and after sleep. Models trained using only medial temporal lobe (MTL) spikes significantly decode concepts before but not after sleep, while models trained using only frontal cortex (FC) spikes decode concepts after but not before sleep. These findings suggest a system-wide distribution of information across neural populations that transforms over wake/sleep cycles[5]. Such decoding of internally generated memories suggests a path towards brain-computer interfaces to treat episodic memory disorders through enhancement or muting of specific memories.

After spinal cord injury (SCI), the "spared" tissue below the lesion contains undamaged cells that could support or augment recovery, but targeting these cells requires a clearer understanding of their injury responses and capacity for repair. Here, we used single nucleus sequencing to profile how each cell type in the lumbar spinal cord changes over time after a thoracic injury. We present an atlas of these dynamic responses and explore two unexpected findings. Amongst neurons, rare cell types expressed a molecular signature of regeneration and amongst microglia, we identified a population of "trauma associated microglia" (TAM). These TAM cells were present in the white matter near degenerating axons and expressed the trophic factors Igf1 and Spp1(OPN). Viral over-expression of Igf1 and Spp1(OPN) expanded the TAM population and promoted the clearance of myelin debris. These findings expose endogenous mechanisms of repair in spared neural tissue, uncovering potential candidates for targeted therapy.

Activation of microglia in the spinal cord following peripheral nerve injury is critical for the development of long-lasting pain hypersensitivity. However, it remains unknown whether distinct microglia subpopulations or states contribute to different stages of pain development and maintenance. We show, using single-cell RNA-sequencing, that nerve injury induces the generation of a male-specific inflammatory microglia subtype, and demonstrate increased proliferation of microglia in males as compared to females. We also show time- and sex-specific transcriptional changes in different microglial subpopulations following injury. Apolipoprotein E (Apoe) is the top upregulated gene in microglia at chronic time points after nerve injury in mice and polymorphisms in the APOE gene in humans are associated with chronic pain. Single-cell analysis of human spinal cord microglia reveals a subpopulation with a disease-related transcriptional signature. Our data provide a detailed analysis of transcriptional states of mouse and human spinal cord microglia, and identify a previously unrecognized role for ApoE in neuropathic pain.

Our earlier study led to the conclusion that human narcolepsy with cataplexy was caused by the loss of forebrain hypocretin (orexin) neurons in the hypothalamus. We now report that humans having narcolepsy with cataplexy also have a 46% decrease in the number and an 18% increase in the size of brainstem neuromelanin-pigmented locus coeruleus (LC) neurons and increased microglial activity in LC. However, no such changes are observed in the LC of hypocretin peptide depleted narcoleptic mice, hypocretin neuron depleted orexin-tTA/TetO-DTA (orexin-DTA) narcoleptic mice or in narcoleptic dogs. Sodium oxybate, an effective treatment for narcolepsy, decreased the size of LC norepinephrine neurons and increased the number and size of LC microglial cells in mice. Our results indicate that the autoimmune process believed to cause human narcolepsy affects both forebrain Hcrt neurons and brainstem LC norepinephrine cells. Addressing both forebrain and brainstem pathologies may improve understanding of, and treatments for, human narcolepsy.

Lesion network mapping (LNM) has emerged as a popular framework to map the network mechanism of brain disorders using lesions and normative brain connectome (NBC)1. It was first demonstrated in neurological symptoms and was rapidly extended to a broad range of brain disorders in the past decade. A recent study by Van den Heuvel et al. questioned the methodological foundations of LNM2. We here raise concerns regarding their errors and biases in the methodology and visualization. The conclusion of that study--LNM maps circumscribed brain changes mostly to one and the same outcome--is not supported by the data presented.

During wake, sound-evoked and spontaneous neural activity of the auditory cortex evolve in distinct subspaces whereas anesthesia disrupts sound responses and merges these spaces. To evaluate if similar modifications of the sound representation geometry explain sensory disconnection during sleep, we followed large neural populations of the mouse auditory cortex across slow wave sleep and wakefulness. We observed that sleep dampens sound responses but preserves the geometry of sound representations which remain separate from spontaneous activity. Moreover, response dampening was strongly coordinated across neurons and varied throughout sleep spanning from fully preserved response patterns to population response failures on a fraction of sound presentations. These failures are more common during high spindle-band activity and more rarely observed in wakefulness. Therefore, in sleep, the auditory system preserves sound feature selectivity up to the cortex for detailed acoustic surveillance, but concurrently implements an intermittent gating mechanism leading to local sensory disconnections.

Histopathological studies of multiple sclerosis (MS), a demyelinating disease of the central nervous system (CNS), and its animal model, experimental autoimmune encephalomyelitis (EAE), have found newly formed leaky vessels in demyelinated acute and chronic plaques, in addition to blood-brain barrier (BBB) damage in existing vessels, that exacerbate disease pathology by increasing infiltration of immune cells. Which vessel subtypes and signaling pathways generate these aberrant vessels is poorly understood. Using single-cell RNA-sequencing and in vivo validation, we find that transcriptome signatures of neo-angiogenesis arise in venous endothelial cells in both acute and chronic EAE, and correlate with upregulation in VEGF-A signaling. These neo-angiogenic markers are also increased in acute and chronic MS lesions. Treatment with a VEGF-A blocking antibody diminishes neo-angiogenic transcriptomic signatures and vascular proliferation in vivo, but does not restore BBB function or ameliorate significantly EAE pathology. Therefore, anti-angiogenic therapies in combination with immunomodulatory therapies may benefit MS progression.

Deciphering the neural code requires identifying its fundamental symbols or code-words. Neural activity is usually interpreted either as a rate code - based on average spike counts - or as a temporal code, which distinguishes patterns with identical counts. Yet, the symbols of the code remain undefined. Here we show that the symbols can be clearly defined by parsing auditory spike trains into variable-duration "packets", within which precise spike timing is irrelevant. Because packets vary in duration, this is not a rate code. A single neuron could encode very different stimuli depending on the number of spikes it produced per packet. The packet-based code enables real-time readout upon packet completion due to its instantaneous code structure and maximizes information capacity at both single-neuron and population levels.

Any learning agent must balance between exploiting its knowledge and exploring new alternatives. Schizophrenia patients are known to have maladaptive exploration-exploitation (E-E) balance1,2 and are impaired at reversal learning tasks as early as their first psychotic episode. The cortico-basal ganglia (BG)-dorsolateral prefrontal cortex (DLPFC) network plays a significant role in learning processes3,4. However, how this network maintains the E-E balance and what alters the balance in schizophrenia remains elusive. Using a combination of extracellular recordings, pharmacological manipulations, macro-stimulation techniques, and an adaptive reinforcement learning model, we show that in the non-human primate (NHP), the external segment of the globus pallidus (GPe, the central nucleus of the BG network) maintain this balance. Furthermore, whereas the chronic, low-dose administration of N-methyl-D-aspartate (NMDA) receptor (NMDA-R) antagonist, phencyclidine (PCP) leads to E-E imbalance, low-frequency GPe macro-stimulation restores it. E-E balance provides a holistic framework to resolve some of the apparent paradoxes that have emerged within schizophrenia research2. Our findings suggest that Schizophrenia symptoms may reflect abnormal DLPFC-BG E-E balance, and GPe stimulation may be advantageous for these patients.

Human brain organisation involves the coordinated expression of thousands of genes. For example, the first principal component (C1) of cortical transcription identifies a hierarchy from sensorimotor to association regions. Here, optimised processing of the Allen Human Brain Atlas revealed two new components of cortical gene expression architecture, C2 and C3, which are distinctively enriched for neuronal, metabolic and immune processes, specific cell-types and cytoarchitecture, and genetic variants associated with intelligence. Using additional datasets (PsychENCODE, Allen Cell Atlas, and BrainSpan), we found that C1-C3 represent generalisable transcriptional programmes that are coordinated within cells, and differentially phased during foetal and postnatal development. Autism spectrum disorder and schizophrenia were specifically associated with C1/C2 and C3, respectively, across neuroimaging, differential expression, and genome-wide association studies. Evidence converged especially in support of C3 as a normative transcriptional programme for adolescent brain development, which can lead to atypical supra-granular cortical connectivity in people at high genetic risk for schizophrenia.

With dense axonal connectivity to every region of cortex, the thalamus plays a central role in the nervous system from sensory processing to cognitive functions. Yet how tissue maturation of the thalamus unfolds during childhood and contributes to typical or atypical development is not clear. Through several large datasets, we provide here a thalamic portrait of fine-scale structural development whose nuclei develop along unique trajectories, some of which diverge from predictions of developmental theory. We find that those thalamic nuclei which show the most protracted development are at the greatest risk for later clinical differences in schizophrenia. The spatial pattern across thalamic nuclei for early psychosis risk is associated with a unique neuroreceptor fingerprint with implications for symptom severity.

Learning where and when rewards like food and water are available is essential for survival1,2. In the simplest cases where resource availability is stable, animals can learn reward contingencies by integrating outcomes across repeated samples of each possible action. In more natural settings, however, reward availability is governed by structured higher-order rules such as depletion and repletion over time. To adapt flexibly to such changing environments, optimal choices require meta-learning wherein animals learn how to learn from external feedback, ultimately enabling them to infer the underlying reward structure from abstract, generalizable rules rather than relying solely on recent outcomes3,4. The existence of meta-learning in animal behavior is well established3-8, yet the neural circuits and computations that implement it remain poorly understood9-11. Here we investigated meta-learning using a spatial foraging task in which rats acquired a depletion-repletion rule that regulated reward availability, and carried out longitudinal, high-density recordings from the medial prefrontal cortex (mPFC). We show that meta-learning engages specific, systematic changes in mPFC neural dynamics that embed the learned rule and thereby alter how the network learns action values from reward outcomes. These dynamics are based on mixed coding of task structure and value in individual mPFC neurons. At the population level, this coding organizes into low-dimensional dynamical motifs that generalize across task conditions. As meta-learning progresses, these motifs are reshaped to instantiate both rule-guided inference of future states before outcome delivery and rule-based value updating during the outcome period. These results indicate that meta-learning sculpts pre-existing prefrontal dynamics to support the acquisition of new, generalizable reward-learning strategies.

Preprints now disseminate a large share of biomedical research before peer review. Because they have not yet passed peer review, some scientists regard preprint claims as unverified or potentially unreliable, yet how much those claims change before publication has so far been quantified only in smaller cohorts, with results that vary by field and topic. Here, we compiled every bioRxiv preprint posted between 2018 and 2025 that we could match by DOI to a peer-reviewed published version, yielding 72,644 preprint-publication pairs. Using a large language model (Claude Sonnet 4.6), we parsed every preprint-publication abstract pair into one primary and two secondary claims, and classified each pair for content change (unchanged, minor, major) and hedging shift (more cautious, more confident, unchanged). On a validation subsample, the model agreed with two independent domain experts about as well as the experts agreed with each other (Cohens kappa 0.63 to 0.66). The primary claim was unchanged in 39.9% of abstracts, minorly revised in 50.0%, and substantially revised in only 10.2%. Hedging shifts were uncommon and asymmetric, with twice as many claims becoming more cautious as more confident (8.4% vs 4.2%). Major revisions were more frequent after long peer review (14.1% in the slowest versus 7.0% in the fastest tertile of review time) and declined over the study period (17.0% in 2019 to 5.7% in 2024). Over the same period, biomedical papers that were never posted as preprints were retracted at roughly twice the rate of those that were. Together, these data show that the move from preprint to peer-reviewed publication leaves the central claims of most biomedical abstracts intact, indicating that preprints are a reliable source of biomedical research.

Chronic pain engages distributed cortical and subcortical circuits, and large-scale intracranial recordings in humans offer a valuable opportunity to characterize its neural signatures. Here, we recorded multi-day stereoelectroencephalography (sEEG) from six participants with refractory chronic neuropathic pain, each implanted with sEEG electrodes spanning dozens of cortical and subcortical structures. Using simultaneous chronic pain ratings, we decoded spontaneous high versus low pain states within individuals (median area under the curve = 0.72; five of six participants performed above chance). Pain-predictive signals were broadly distributed and highly participant-specific. However, mapping the spatial distribution of pain-predictive features revealed preferential representation within canonical macroscale networks: beta-band activity in the default mode network and high-gamma activity in the salience network. These results demonstrate that intracranial recordings can capture distributed, network-organized representations of spontaneous chronic pain states.

The ability to estimate the value associated with a specific stimulus or action is essential for adaptive behavior. Value can be updated either incrementally through experience or rapidly by inference based on latent environmental structure. Yet, how the brain implements and transitions between these modes of value computation remains unclear. To address this question, we examined the neuronal mechanisms underlying reversal learning. Mice were trained in an odor-outcome association task either with stable contingencies or with dynamically changing contingencies. Mice trained on stable contingencies formed long-term value representations that depended on synaptic plasticity in the basolateral amygdala (BLA). In contrast, mice exposed to repeated reversals acquired the ability to infer values, independent from plasticity in BLA, enabling faster learning but with more rapid memory decay. Recurrent neural network models (RNNs) trained with continuous weight updates recapitulated this transition, shifting from plasticity-based to dynamics-based value computation. Neural activity in the BLA encoded both value and contextual information necessary for computing value based on latent task structure, similar to those found in the RNNs. Disrupting BLA activity before cue delivery preferentially impaired dynamics-based value updating. Furthermore, mice could learn distinct correlation structures that enabled structure-specific value inference. Together, these findings provide a mechanistic framework for fast value updates via inference, a core feature of intelligent behavior.

The mesocortical dopamine system is comprised of midbrain dopamine neurons that predominantly innervate the medial prefrontal cortex (mPFC) and exert a powerful neuromodulatory influence over this region1,2. mPFC dopamine activity is thought to be critical for fundamental neurobiological processes including valence coding and decision-making3,4. Despite enduring interest in this pathway, the stimuli and conditions that engage mPFC dopamine release have remained enigmatic due to inherent limitations in conventional methods for dopamine monitoring which have prevented real-time in vivo observation5. Here, using a fluorescent dopamine sensor enabling time-resolved recordings of cortical dopamine activity in freely behaving mice, we reveal the coding properties of this system and demonstrate that mPFC dopamine dynamics conform to a selective attention signal. Contrary to the long-standing theory that mPFC dopamine release preferentially encodes aversive and stressful events6-8, we observed robust dopamine responses to both appetitive and aversive stimuli which dissipated with increasing familiarity irrespective of stimulus intensity. We found that mPFC dopamine does not evolve as a function of learning but displays striking temporal precedence with second-to-second changes in behavioral engagement, suggesting a role in allocation of attentional resources. Systematic manipulation of attentional demand revealed that quieting of mPFC dopamine signals the allocation of attentional resources towards an expected event which, upon detection triggers a sharp dopamine transient marking the transition from decision-making to action. The proposed role of mPFC dopamine as a selective attention signal is the first model based on direct observation of time-resolved dopamine dynamics and reconciles decades of competing theories.