bioRxiv Subject Collection: Neuroscience's Journal

Metabolic Flexibility of Microglia: Energy Substrate Utilization and Impact on Neuronal Metabolism
Microglia, the main resident immune cells of the brain, play critical roles in maintaining neuronal function and homeostasis. Microglia metabolic flexibility enables rapid adaptation to environmental changes, yet the full extent of their metabolic capabilities and influence on neuronal metabolism remains unclear. While microglia predominantly rely on glucose oxidative metabolism under homeostatic conditions, they shift toward glycolysis upon proinflammatory activation. In this study, we investigated microglial metabolism and its impact on neuronal metabolic homeostasis using isotope tracing with stable carbon 13C-enriched substrates and gas chromatography-mass spectrometry (GC-MS) analysis. Primary microglia were incubated with 13C-labeled glucose, glutamine, or GABA in the presence or absence of lipopolysaccharide (LPS) to assess metabolic adaptations upon an inflammatory challenge. Additionally, neurons co-cultured with quiescent or activated microglia (either with LPS or amyloid-{beta}) were incubated with 13C-enriched glucose to examine microglia-neuron metabolic interactions. Our findings confirm that microglia readily metabolize glucose and glutamine, with LPS stimulation slightly changing the glycolytic activity, as indicated by subtle changes in extracellular lactate. Importantly, we demonstrate for the first time that microglia take up and metabolize the inhibitory neurotransmitter GABA, suggesting a novel metabolic function. Furthermore, microglial presence directly influences neuronal metabolism and neurotransmitter homeostasis, highlighting a previously unrecognized aspect of neuron-microglia metabolic crosstalk. Collectively, these findings provide new insights into microglial metabolism and its role in neuronal function, with implications for neuroinflammatory and neurodegenerative diseases in which microglial metabolism is dysregulated.

Synaptic and neural pathway redundancy enables the robustness of a sensory-motor reflex and promotes predation escape in C. elegans
As a basic unit of the nervous system, the sensory-motor reflex circuit is fast and robust. However, it is not entirely clear how this robustness is achieved, given that various genetic perturbations can disrupt the function of the sensory neurons. By mapping the molecular basis of neuronal connections in the touch response circuit of Caenorhabditis elegans, we found prevalent genetic redundancy at neural pathway, synaptic, and molecular levels, which ensures that sensory signals can be relayed to command interneurons that control motor output. We also discovered developmental remodeling of the anterior circuit, which leads to the pruning of larval synapses, establishment of a second pathway that activates additional interneurons, and lateralization of the circuit. Finally, we found that the synapses that appeared to be functionally redundant in a simple touch assay contribute to the extent of reversal response in an additive manner, which may help the organism escape from predators.

Language network functional connectivity in infancy predicts developmental language trajectories
Although developmental language delays affect approximately 10% of children in the general population, the neurodevelopmental mechanisms that support normative language acquisition, and atypicalities that may predict later language delay, across the first year of life are poorly understood. Here, resting-state fMRI data from the Baby Connectome Project was used to evaluate age-related changes in language network functional connectivity and alterations associated with suboptimal language development. Additionally, a data-driven machine learning algorithm was used to partition our sample into three groups who showed Delayed, Typical, and Advanced trajectories of language development. These groups reliably differed on several assessments of language ability during infancy and toddlerhood. Using a priori brain regions involved in adult language processing, a seed-based functional connectivity analysis showed broad age-related increases in functional synchrony and specialization throughout the infant language network. Additionally, the Delayed group showed several atypical patterns of functional connectivity with language regions. Importantly, the magnitude of connectivity differences consistently predicted later language scores at two-year outcome across several different language assessments. These findings add to our understanding of normative neurodevelopmental patterns underlying language acquisition, and identify several potential biomarkers associated with language delay that could serve as future targets to inform diagnoses and clinical interventions.

What one sees depends on how far the eye has moved
Humans explore visual scenes through frequent, rapid gaze shifts known as saccades. These movements redirect the high-acuity region of the retina toward objects of interest, thus selecting information based on location. Here, we show that saccade amplitude provides a separate and complementary form of selection, effectively filtering visual information by spatial frequency rather than location. Specifically, a reduction in saccade amplitude attenuates post-saccadic visual sensitivity in an amplitude-dependent range of low spatial frequencies. This effect is highly robust, so that even minute changes in saccade size considerably affect visibility. We show that this phenomenon arises from the way the magnitude-dependent kinematic characteristics of saccades transform the visual world into a spatiotemporal flow: post-saccadic visibility closely follows theoretical predictions based on the spatial information that saccade transients convey within the temporal bandwidth of retinal sensitivity. Thus, saccades not only guide selection based on location, but also filter visual information based on content, actively shaping perception.

Age associations with cortical and subcortical brain structure in adolescents age 9-17
Introduction: Adolescence is a pivotal period in brain structural development and maturation. However, investigation of cortical and subcortical brain changes during this time have been limited by small sample size and have generally examined the brain at the level of predetermined regions of interest. The recently developed Fast Efficient Mixed-Effects Algorithm (FEMA) allows for increased computational speed using mixed-effects models applied at the voxel or vertex level, as well as across multiple regions of interest. Methods: We extended the existing FEMA framework to represent predictors using natural spline basis functions, enabling us to model nonlinear trajectories of brain structure as a function of age. We then applied this model to the The Adolescent Brain Cognitive DevelopmentSM Study (22,651 observations from 10,521 unique subjects aged 9.00-17.77) to study the age-related trajectories of tabulated cortical and subcortical volumes, vertexwise cortical thickness and surface area, and voxelwise volume assessed using the Jacobian. Models are reported separately in males and females. Results: Global volume variables, including total subcortical gray matter volume, peaked near 13 years in females and 15 years in males. Vertexwise cortical surface area followed an inverted U-shaped curve, whereas vertexwise cortical thickness followed a monotonic decrease during the age range studied. Voxelwise imaging analysis revealed regional differences in age trajectories at the subregional level. Discussion: The results of this work replicate and extend prior findings related to adolescent brain development, and illustrate distinct spatiotemporal patterns of structural changes in subcortical regions. The updated FEMA framework is publicly available for use in similar large datasets.

The engulfment receptor Draper is required for epidermal dendrite ensheathment
Our experience of the external world is shaped by somatosensory neurons (SSNs) that innervate our skin and mediate responses to a range of environmental stimuli. The precise innervation patterns and response properties of SSNs are determined in part by specialized interactions with resident skin cells. One such interaction involves the preferential ensheathment of some SSNs by epidermal cells, an evolutionarily conserved intercellular interaction that regulates SSN morphogenesis and mechanical nociceptive sensitivity in Drosophila. The morphogenetic events during ensheathment resemble phagocytic engulfment, therefore we hypothesized that phagocytic receptors mediate molecular recognition of neurites to induce ensheathment. From a screen of epidermally expressed phagocytic receptors we found that the nimrod receptor gene Draper (Drpr) functions in epidermal cells to promote ensheathment. Endogenous Drpr accumulates at sites of epidermal ensheathment but not at epidermal contact sites with unensheathed neurites. Furthermore, overexpressing Drpr increased ensheathment selectively on neurons that are normally ensheathed, suggesting that molecular recognition by Drpr accounts for the specificity of ensheathment. Indeed, we found that an extracellular reporter for the Drpr ligand Phosphatidylserine (PS) accumulates at sites of ensheathment, and that preventing extracellular PS exposure by overexpressing the PS Flippase ATP8a blocked ensheathment. We additionally found that Orion, which encodes a chemokine-like protein that bridges Drpr-PS interactions, is required for sheath formation. Finally, we found that increasing ensheathment by overexpressing Drpr enhanced nociceptor sensitivity to mechanical stimulus. Altogether, these studies show that Drpr acts in epidermal cells to mediate molecular recognition events that drive ensheathment of neurites marked by extracellular PS.

Model-based and model-free valuation signals in the human brain vary markedly in their relationship to individual differences in behavioral control
Human action selection under reinforcement is thought to rely on two distinct strategies: model-free and model-based reinforcement learning. While behavior in sequential decision-making tasks often reflects a mixture of both, the neural basis of individual differences in their expression remains unclear. To investigate this, we conducted a large-scale fMRI study with 179 participants performing a variant of the two-step task. Using both cluster-defined subgroups and computational parameter estimates, we found that the ventromedial prefrontal cortex encodes model-based and model-free value signals differently depending on individual strategy use. Model-based value signals were strongly linked to the degree of model-based behavioral reliance, whereas model-free signals appeared regardless of model-free behavioral influence. Leveraging the large sample, we also addressed a longstanding debate about whether model-based knowledge is incorporated into reward prediction errors or if such signals are purely model-free. Surprisingly, ventral striatum prediction error activity was better explained by model-based computations, while a middle caudate error signal was more aligned with model-free learning. Moreover, individuals lacking both model-based behavior and model-based neural signals exhibited impaired state prediction errors, suggesting a difficulty in building or updating their internal model of the environment. These findings indicate that model-free signals are ubiquitous across individuals, even in those not behaviorally relying on model-free strategies, while model-based representations appear only in those individuals utilizing such a strategy at the behavioral level, the absence of which may depend in part on underlying difficulties in forming accurate model-based predictions.

NKCC1 a Regulator of Glioblastoma Progression
Glioblastoma (GBM) is the most common and aggressive primary brain tumor in adults, with poor prognosis despite multimodal therapy. Chloride cotransporters NKCC1 and KCC2 are key regulators of intracellular chloride levels and thereby determine whether GABA acts inhibitory or excitatory. In GBM, disrupted chloride homeostasis promotes proliferation, migration, and stem-like properties, but its clinical relevance is not fully understood. We analysed NKCC1 and KCC2 expression in glioblastoma samples, considering clinical parameters such as age, gender, and MGMT promoter methylation. Statistical analyses included ROC-based cutoff determination, Kaplan-Meier survival analysis and subgroup. Immunohistochemistry was performed to identify cell types expressing NKCC1. NKCC1 expression was significantly higher in older patients and emerged as a prognostic marker for recurrence-free survival, with lower levels correlating with delayed recurrence, although overall survival was unaffected. NKCC1 was expressed in stem-like, astrocytic, and progenitor cells, but not in mature neurons. These findings identify NKCC1 as a regulator of GBM progression and recurrence, linking chloride transporter imbalance to GABAergic signalling. Targeting NKCC1 and restoring chloride homeostasis may provide promising new treatment strategies.

Overlooked neuroanatomical markers of face processing and developmental prosopagnosia in posteromedial cortex
Recent functional imaging studies implicate human posteromedial cortex (PMC)--composed of the posterior cingulate cortex, the precuneus, and retrosplenial cortex--in face processing. Separately, anatomical studies have identified previously overlooked cortical folds (sulci) in PMC associated with higher-level cognitive abilities. Here, we tested whether these newly identified sulci support face processing in neurotypical individuals and individuals with developmental prosopagnosia (DP). After manually labeling 1,642 sulci in 164 hemispheres, we first identified a gradient of face selectivity along the anterior-posterior axis of PMC that was consistent across three samples, including DP individuals. Second, we discovered a new anatomical locus in PMC that differed structurally and functionally between neurotypical and DP individuals. Finally, data-driven analysis revealed that right-hemisphere PMC sulcal morphology was associated with face recognition ability. These findings reveal a sulcal network in PMC that supports face processing, and they identify the first structural neuroanatomical marker of face processing deficits in PMC.

A mouse model for cerebral/cortical visual impairment (CVI) impairs vision and disrupts the spatial frequency tuning of neurons in visual cortex
Cerebral/cortical visual impairment (CVI) is a visual disorder associated with perinatal hypoxic injury. The pathophysiology of CVI is poorly understood in part because of the lack of an animal model. Here we developed a murine model of CVI from existing rodent early postnatal hypoxia models for periventricular leukomalacia. Exposure to hypoxia during the equivalent to the human third trimester did not perturb motor function but caused severe impairments in binocular depth perception and visual acuity. Impaired vision was associated with normal retinal function, but reduced size of the visual thalamus, and aberrant tuning for spatial frequency by populations of excitatory neurons in primary visual cortex. This murine model of CVI provides a framework for triangulating circuit deficits with specific visual impairments and testing potential therapeutic interventions.

Loss of Brain Structural qMRI Signatures in Aging
The human brain undergoes significant changes during aging, affecting its macrostructure and microstructure features and impacting its functionality. These aging related changes are oftentimes linked to different neurodegenerative diseases. Previous studies using MRI and molecular analyses have proposed that aging in each individual brain leads to increased similarity between different brain regions, whereas on the population level, they show greater variability. We investigated these hypotheses using multi-parametric, microstructural, quantitative MRI (qMRI) approaches. Using seven qMRI parametric methods in data from young and older adults, we created unique microstructural signatures of 128 brain regions. Our results supported both hypotheses: First, that inter-regional microstructural similarity is stronger within older individuals; and second, that the variability in microstructural signatures across individuals is also greater in older than in younger adults. Our findings provide new insights into the brain's microstructural changes during aging and demonstrate the potential of using multi-parameter qMRI techniques in neuroscience research.

Hyperspherical geometry positions the lipidome as a partly independent axis of human brain organization
The brain's transcriptome is well mapped, but the spatial organization of lipids -- over half of brain dry mass -- remains poorly defined. We profiled lipidomic (419 species) and transcriptomic (15,013 genes) signatures from 35 anatomically defined regions in four healthy adult donors, measured from the same tissue samples. Both modalities recapitulate major neuroanatomical divisions, yet the lipidome shows distinctive features: a pronounced white-gray asymmetry, a smooth neocortical rostrocaudal gradient, and limbic-specific lipid clusters absent in transcriptomic space. Using regression against gene-expression principal components and two null models (random and anatomy-aware), we define three data-driven classes of lipid-gene relationships: Synchronizers (10%) tightly coupled to gene programs, Anchors (67%) predictable at levels expected from gross anatomy and cell-type composition, and Drifters (23%) largely transcription-independent. A simple geometric framework unifies these patterns: after centering and normalization, molecular profiles lie on a hypersphere where two interpretable coordinates -- polar latitude relative to transcriptome-defined white/gray poles and nearest-gene angular distance -- jointly index coupling; an analytical null quantitatively explains the observed scaling. Key effects are robust in leave-one-donor-out analyses and position the lipidome as a partly independent organizational axis of the human brain. Broadly, our results provide an interpretable geometric framework for multi-omics integration, supported by an interactive platform.

The trimeric structures of the extracellular domains of FAM171A1 and FAM171A2 neuronal proteins belong to a novel structural superfamily
Cell surface molecules play fundamental roles in cell-cell communication, attraction, or repulsion, and when expressed in neurons they are often implicated in neurological disorders. FAM171 is a family of three type-I transmembrane domain cell surface proteins (FAM171A1, FAM171A2, and FAM171B) expressed in several human tissues and especially enriched in the brain. Recent findings suggest that FAM171A1 transduces signals between the cell surface and the cytoskeleton. Genetic evidence links FAM171A1 to multiple cancers and FAM171A2 to neurodegenerative diseases, including Alzheimers and Parkinsons diseases. Despite multiple connections with severe human diseases, no information is currently available on their neuronal expression, their structure, or oligomerization. Here we show that, structurally, the monomeric ectodomains of human FAM171A1 and FAM171A2 has a new architecture with a novel combination of two domains. Furthermore, their ectodomains oligomerize to form an equilateral trimer. In addition, the ectodomain of FAM171A1 has the propensity to form larger trimer-trimer assemblies at high concentrations. In the adult mouse brain, Fam171a1 and Fam171a2 mRNAs are broadly distributed in the hippocampus, with predominant expression in excitatory neurons and additional expression in inhibitory neurons. At the protein level, mouse Fam171a1 and Fam171a2 are mainly localized to the hippocampal neuropil, suggesting localization to neuronal processes. Together, these results provide novel insights into the structure and oligomerization of the extracellular domain of FAM171A1 and FAM171A2, suggesting important roles in ligand binding and signaling.

Screening of a kinase library in human Huntington disease iPSC derived striatal precursor neurons reveals a neuroprotective effect of PKC alpha and PKC beta1 inhibition
The loss of striatal medium spiny neurons is a hallmark of Huntington's disease (HD). To identify potential disease-modifying treatments, we previously developed a human neuronal model by immortalizing and differentiating HD patient-derived iPSCs into highly homogeneous striatal precursor neurons (ISPNs). Using a 96-well screening platform, and two rounds of re-screening, we tested a kinase inhibitor library and identified 5 compounds that protected HD ISPNs from mutant huntingtin (mHTT)-induced toxicity. Among these, we prioritized the PKC-/{beta}1 inhibitor GO6976, which rescued HD ISPNs from mHTT toxicity in a dose-dependent manner. Further, we found increased phosphorylation of PKC- and PKC-{beta}1 in HD cells and tissues, while their overexpression was toxic to HD ISPNs. Knockdown of PKC-/{beta}1 protected the neurons, and both isoforms interacted and colocalized with HTT. These results suggest that PKC-/{beta}1 plays a role in HD neurodegeneration, and that inhibiting their activity may offer a potential therapeutic approach for HD.

The nuclear receptor NR4A1 serves as a neutrophil-intrinsic regulator mitigating stroke severity
Ischemic stroke is accompanied by recruitment and activation of immune cells which play an important role in the progression of the brain damage. The nuclear receptor NR4A1 emerged as a key regulator within the inflammatory response of several immune diseases by regulating immune cell activation. In this study, we investigated the role of NR4A1 in the activation and recruitment of brain resident and peripheral immune cells after cerebral ischemia. Here, we show that NR4A1 mediates an anti-inflammatory and damage-limiting effect after stroke. This effect is largely mediated by neutrophil recruitment and importantly, NR4A1 activation with its ligand Cytosporone B improves functional outcome and reduces brain damage. Modulation of NR4A1 is therefore a promising therapeutic target for the treatment of the nuclear receptor NR4A1 in the activation and recruitment of peripheral and brain resident immune cells after cerebral ischemia and its consequences for stroke outcome. We demonstrate that NR4A1 ablation augments neutrophil activation and CNS recruitment within days after stroke thereby increasing infarct size, CNS inflammation, neuronal damage and deteriorating functional outcome. This effect is mediated via modulation of cell-intrinsic neutrophil function and maturation as illustrated by neutrophil-specific NR4A1 ablation and mixed bone-marrow chimera experiments. Notably, the NR4A1 agonist Cytosporone B reduced CNS neutrophil infiltration, infarct size and functional outcome after stroke in a bicentric preclinical stroke trial, demonstrating that NR4A1-mediated control of neutrophil reactivity is amenable to pharmacological modulation. In humans, NR4A1 expressing neutrophils are present in the peripheral blood of stroke patients and neutrophil NR4A1 expression correlates with improved long-term outcome after 3 months. Furthermore, NR4A1 expression in brain parenchyma neutrophils is negatively correlated with neuronal cell loss, illustrating a role of NR4A1 in regulating neutrophil mediated neuronal cell death in human stroke. Together our data reveal the nuclear factor NR4A1 as a brake of intrinsic neutrophil activity controlling neutrophil-mediated brain inflammation and neurotoxicity in stroke which may serve as a novel therapeutic target to limit inflammation-associated augmentation of ischemic damage after stroke.

Noradrenergic inputs to the basolateral amygdala have bidirectional effects on coping behavior and neuronal activity in mice
Norepinephrine (NE) signaling is disrupted in stress disorders, with insufficient NE signaling implicated in major depressive disorder and hyperactive NE signaling associated with post-traumatic stress disorder, suggesting that adequate mood regulation requires optimal NE levels. The basolateral amygdala (BLA) is a hub for stress processing and receives dense noradrenergic innervation from the locus coeruleus (LC), the primary noradrenergic nucleus in the brain. The relationship between LC activity and cognitive/behavioral function during fear conditioning has been described as an inverted U, in which moderate LC activity, and subsequent NE release, is required for adaptive coping to threats, while hyperactive LC-NE signaling drives maladaptive behavioral responses. We used fiber photometry to measure NE signaling in the mouse BLA during acute behavioral responses to escapable and inescapable stressors, and then used an optogenetic approach to stimulate the noradrenergic terminals in the BLA at different frequencies to evaluate effects on coping behavior and cFos expression in the LC-BLA circuit. We found that low-frequency stimulation of the circuit inhibited both passive coping and BLA neuronal activity, while high-frequency stimulation had the opposite effect; the behavioral effects were not mediated by sex, but the cFos effects were specific to males. This study represents an expansion of the inverted U framework to encompass LC-BLA signaling driving acute behavioral responses to stress.

Characterization of adult hippocampal neurogenesis in the novel AppSAA Knock-in Alzheimer's disease mouse model
Adult hippocampal neurogenesis (AHN) declines with age and is thought to be severely exacerbated in neurodegenerative disorders like Alzheimer's disease (AD). Despite numerous efforts to understand how AHN is altered in AD mouse models, results have been inconsistent, largely due to limitations of first-generation transgenic AD mouse models. The newly developed App Knock-in models address many of these limitations. Here, we provide the first in-depth characterization of hippocampal cell populations and AHN across different ages in the novel AppSAAKnock-in (AppKI) mouse model. Our findings reveal that AppKI mice show no early deficits in excitatory dentate granule cells or inhibitory interneurons at 2 and 4 months, but significant loss of both populations emerges by 6 months of age. We also identified a progressive decline in AHN with the survival of newborn neurons being impaired first at 4 months, followed by a deficit in proliferation at 6 months. Furthermore, we demonstrate that exposure to enriched environment, a form of hippocampus-engaged exploration, robustly enhances AHN in AppKI mice, primarily by promoting survival. In conclusion, our study provides a foundational characterization of the AppKI model, establishing a timeline for cellular and AHN deficits.

Ineffective behavioral rescue despite partial brain Dp427 restoration by AAV9-U7-mediated exon 51 skipping in mdx52 mice
The mdx52 mouse model exhibits a common mutation profile associated with brain involvement in Duchenne muscular dystrophy (DMD), characterized by heightened anxiety, fearfulness, and impaired associative fear learning. Deletion of exon 52 disrupts the expression of two dystrophins found in the brain (Dp427 and Dp140), and is eligible for therapeutic exon-skipping strategies. We previously demonstrated that a single intracerebroventricular administration of an antisense oligonucleotide (ASO) targeting exon 51 of the Dmd gene could restore 5% to 15% of Dp427 expression. This treatment reduced anxiety and unconditioned fear in mdx52 mice, improved fear conditioning acquisition, and partially improved fear memory tested 24 hours later. To improve the restoration of Dp427 and induce a long-lasting therapeutic effect, we employed a vectorized approach using an AAV-U7snRNA vector to deliver antisense sequences to the brains of mdx52 mice. We evaluated two AAV serotypes known for their brain transduction efficiency (AAV9 and RH10) and two delivery routes, intracisterna magna and intracerebroventricular (ICV) injections, to maximize brain targeting. Based on GFP expression data, we selected the AAV9 capsid and a bilateral ICV delivery route. Using this approach, we demonstrated that ICV administration of AAV9-U7-Ex51M induced exon 51 skipping and restored Dp427 expression in the brains of adult mdx52 mice, though with significant variability among individuals. While a few mice showed high Dp427 expression levels, the average restoration was limited to approximately 6% to 12%. In conclusion, inducing exon skipping in the brains of adult mdx52 mice using the vectorized AAV9-U7 approach was less effective than synthetic ASO treatment and did not improve the emotional behavior of mdx52 mice.

A multi-omic atlas of human autonomic and sensory ganglia implicates cell types in peripheral neuropathies
The human peripheral nervous system (PNS) consists of many ganglia, including sympathetic ganglia (SG) and dorsal root ganglia (DRG), that house the cell bodies of many constituent neuron types and non-neuronal cells of the PNS. However, the molecular and cellular diversity of these human PNS cell types and their implications in human diseases remain elusive. By generating an integrated single-cell multi-omic atlas of human SG and DRG, we provide comprehensive transcriptional and epigenomic landscapes of various cell types in these peripheral ganglia. While the major cell types and their transcriptional and epigenomic features are similar between human SG and DRG, we identify key transcriptional and epigenomic differences between SG and DRG cell types, highlighting the distinct molecular mechanisms underlying their specific functions. Moreover, by mapping the expression and chromatin accessibility of disease-associated genes in human SG and DRG cell types, we identify cellular and molecular mechanisms that may underlie various peripheral neuropathies. This atlas serves as a valuable resource for understanding the intricate cell-type-specific molecules and interactions in the human PNS and their implications in human health and diseases.

Temporal evolution and spatial heterogeneity of cerebral cortical-depth profiles of the BOLD-fMRI response
Functional MRI measures brain activity by tracking the associated hemodynamic response, which is shaped by the vascular anatomy. Microscopy studies have shown that the hemodynamic response initiates within the cerebral cortex then spreads upwards to the pial surface, where the largest fMRI signal changes are seen, however responses at the cortical surface exhibit poor neuronal specificity. Motived by this, we characterized the time-evolution of the fMRI response in humans. At some cortical locations the fMRI response peaked at the pial surface but in others it peaked within the cortex, likely due to the sparsity of large pial vessels and the dense capillary bed in middle cortical layers. We observed the earliest response onset in middle cortical depths, which is also the site of thalamocortical input. Standard approaches that aggregate fMRI responses over space and time may therefore lose meaningful information, and "physiologically-informed" strategies may enhance neuronal specificity of fMRI measurements.

Optimizing Language Model Embeddings to Voxel Activity Improves Brain Activity Predictions
Recent studies have shown that contextual semantic embeddings from language models can accurately predict human brain activity during language processing. However, most studies use contextual embeddings with the same context length and model layer for all voxels, potentially overlooking meaningful variations across the brain. In this study, we investigate whether optimizing contextual embeddings for individual voxels improves their ability to predict brain activity during reading. We optimize embeddings for each voxel by selecting the best-predicting context length, model layer, or both. We perform this optimization with two different types of stimuli (isolated sentences and narratives), and quantify the performance gains of optimized embeddings over standard fixed embeddings. Our results show that voxel-specific optimization substantially improves the prediction accuracy of contextual semantic embeddings. These findings demonstrate that voxel-specific contextual tuning provides a more accurate and nuanced account of how the contextual semantic information is represented across the cortex.

Differential locus coeruleus-hippocampus interactions during offline states
Patterns of locus coeruleus (LC) activity and norepinephrine (NE) release during non-rapid-eye-movement (NREM) sleep suggest a critical role for the LC-NE system in offline modulation of forebrain circuits. NE transmission promotes synaptic plasticity and is required for memory consolidation, but the field has only begun to uncover how LC activity contributes to coordinated forebrain network dynamics. Hippocampal ripples, a hallmark of memory replay, are temporally coupled with thalamocortical oscillations; however, the circuit mechanisms underlying systems-level consolidation across larger brain networks remain incompletely understood. Here, using multi-site electrophysiology, we examined LC firing in relation to hippocampal ripples in freely behaving rats. LC activity and ripple occurrence were state-dependent and inversely related: heightened arousal was associated with increased LC firing and reduced ripple rates. At finer timescales, LC spiking decreased {approx}1-2 seconds before ripple onset, with the strongest modulation during awake ripples but minimal change during ripple-spindle coupling. These findings reveal state-dependent dynamics of LC-hippocampal interactions, positioning the LC as a key component of a cortical-subcortical network supporting systems-level memory consolidation.

A human spiking computational model to explore sound localization
Sound localization relies on precise processing of binaural cues in medial (MSO) and lateral superior olive (LSO). However, key questions remain on how these two nuclei perform their specific computations depending on sound frequency, the relative contributions of interaural time differences (ITDs) and interaural level differences (ILDs), as well as the role of inhibitory timings. Experimental studies have struggled to address these issues because of technical challenges and a lack of methodological consistency. Here, we present a comprehensive computational model of auditory peripheral and brainstem neural populations to investigate how ITDs and ILDs are encoded by LSO and MSO. We developed a spiking neural network with realistic tonotopic organization and biologically consistent synaptic connections. We tested responses to pure tones and white noise from different locations under three cue conditions: human-recorded head-related transfer functions, isolated ITDs, and isolated ILDs. LSO neurons showed realistic ipsilateral-preferring responses across different stimuli, with cue dependency varying by tone frequency, while white noise responses were driven mostly by ILDs. MSO responses showed heterogeneous tuning, with contralateral preference for low-frequency tones, which got lost for higher-frequency ones, and white noise response driven by ITDs. We further validated two experimental findings: (1) removing MSO inhibition abolished contralateral tuning, and (2) varying the timing between excitation and inhibition produced large shifts in tuning, highlighting the importance of synaptic timing for ITD coding. This model serves as an in silico testbed for auditory research, offering new insights into the functioning of human spatial hearing.

GPT-4 accurately predicts human emotions and their neural correlates
Recent advances in multimodal large language models ((M)LLMs), such as GPT-4, enable them to accurately analyze and describe complex visual scenes, raising the question whether LLMs can also predict human emotional experiences evoked by similar scenes. Here we asked GPT-4 and humans (N = 519) to provide self-reports of 48 unipolar emotions and affective dimensions for emotionally evocative videos and images. We evaluated GPT-4's emotion ratings using three natural socio-emotional stimulus datasets: two video datasets (234 and 120 videos) and one image dataset (300 images). We found that GPT-4 can predict emotions of human observers with high accuracy. The multivariate emotion structure (correlation matrices of emotions' ratings) converged between GPT-4 and humans and across datasets indicating that GPT-4 ratings for different emotions follow similar structural representations as the human evaluations. Finally, we modeled the brain's hemodynamic responses for emotions elicited by videos or images in two fMRI datasets (N = 97) with GPT-4 or human-based emotional evaluations to highlight the usefulness of GPT-4 in neuroscientific research. The results showed that the brain's emotion circuits can be mapped with high accuracy using GPT-4 emotion ratings as the stimulation model. In conclusion, GPT-4 can predict human emotion ratings to the extent that GPT-4 ratings can also model the associated neural responses. Our results indicate that LLMs provide novel and scalable tools that have broad potential in emotion research, cognitive and affective neuroscience, and that it can also have practical applications.

Metabolic and transcriptional adaptations to phagocytosis in microglia sustain their functionality and regenerative properties
Phagocytosis of apoptotic cells, or efferocytosis, is a tightly regulated process that ensures tissue homeostasis and prevents mounting inflammatory responses. In the brain parenchyma, it is executed by microglia, which are encumbered by large numbers of apoptotic debris generated during development, in adult neurogenic niches, aging, and brain diseases. Emerging evidence suggest that phagocytosis is not limited to garbage disposal, but triggers adaptations in the phagocytes that may have a functional impact. To test it we developed an in vivo model of superphagocytosis induced by low cranial irradiation (LCI, 2Gy) that specifically induced apoptosis in the neurogenic niche of the adult hippocampus, synchronizing microglia in a phagocytic state within 6h and leading to full clearance by 24h. Single cell RNA sequencing and metabolomics revealed an unexpected oxidative stress in post-phagocytic microglia, accompanied by catabolic shutdown, mitochondrial remodeling, increased expression of galectin 3, and production of polyamines that led to cell death and compensatory proliferation. To test whether these changes impaired subsequent microglial phagocytosis, we used a glioblastoma model treated with sequential irradiation to induce tumor cell apoptosis. The phagocytosis efficiency of tumor-associated microglia/macrophages was comparable in the first and second apoptotic challenge, suggesting that the metabolic remodeling induced by phagocytosis was adaptive and destined to sustain their functionality. Finally, we assessed the functional impact of post-phagocytosis adaptations using galectin 3 deficient mice under LCI. We found that the recovery of the neurogenic niche after LCI strongly depended on galectin 3, demonstrating the regenerative capacity of post-phagocytic microglia. Overall, our data unveils the complexity of post-phagocytosis adaptations in microglia, underscoring their unexplored therapeutic potential in brain disorders.

Residual Foveal Motion Facilitates Processing of Visually Tracked Objects
Humans and other species visually track moving objects via pursuit eye movements. These movements fail to stabilize the stimulus on the retina, an effect often attributed to errors in oculomotor control. However, it remains unclear whether the resulting residual retinal motion serves visual functions. Using high-resolution eye tracking, we reconstructed foveal motion during discrimination of both stationary and moving stimuli. We show that the retinal motion elicited by pursuit is heavily constrained and strikingly mirrors the motion present during normal active fixation. We then show that this motion performs an important information-compression computational function by equalizing luminance modulations in a low-spatial-frequency range during tracking of natural stimuli. Finally, we show that the resulting visual signals during tracking contribute to perceptual judgments, shifting spatial sensitivity toward lower frequencies relative to fixation. These results suggest that retinal motion during pursuit constitutes an active strategy for encoding space in the joint space-time domain.

Transcranial Alternating Current Stimulation can disrupt or reestablish neural entrainment in a primate model of Parkinson's disease
Transcranial alternating current stimulation (tACS) is a non-invasive brain stimulation method which can affect brain oscillations by inducing neuronal entrainment through modulation of spike timings. tACS has high potential for clinical applications in many neurological disorders in which the stimulation can be delivered to modulate oscillatory activity and disrupt pathological brain oscillations. For instance, in Parkinson's Disease (PD), electrophysiological activity in the motor network often exhibits excessive and hyper-synchronized beta oscillations. However, the development of tACS as a therapeutic intervention for pathological oscillations requires the prior establishment of physiologically effective stimulation parameters. We recorded neuronal activity in the motor cortical area of three parkinsonian non-human primates and examined the influence of tACS-induced electric fields on neural firing patterns. We found that weak extracellular electric fields first disrupt beta-band spike timing patterns by changing the preferred spiking phase of neurons but eventually reestablish neural entrainment with altered phase preferences when electric fields are high. Additionally, we show that frequency-matched stimulation, when stimulation frequency corresponds to endogenous oscillatory activity, significantly enhances neural entrainment. Thus, tACS exhibits significant potential for controlling and modulating pathological oscillatory patterns in many neurological disorders such as PD.

Task-Based Spinal fMRI using a Median Nerve Stimulation Reveals Functional Alterations in Degenerative Cervical Myelopathy
Functional spinal cord MRI (fMRI) is an emerging technique for evaluating sensorimotor responses in both healthy and disease states. However, degenerative cervical myelopathy (DCM), a non-traumatic, age-related pathology characterized by spinal cord compression, often leads to motor and sensory impairments that make standard task-based fMRI paradigms challenging to perform or execute. To overcome this, we implemented an fMRI paradigm using direct electrical stimulation to the median nerve to evoke neural responses in both motor and sub-motor spinal cord pathways. In this study, 23 DCM and 23 aged-matched healthy controls (HC) underwent two task-based spinal cord fMRI sessions with direct, median nerve stimulation to the right upper limb. Motor thresholds were determined on an individual basis by a thumb/finger twitching, and sub-motor thresholds were set to be 15% underneath motor with no visible twitching. Blood Oxygenation Level Depended (BOLD) signals were analyzed across the C5-C8 cervical spinal cord segments. HCs demonstrated increased activation during motor vs sub-motor stimulation (1155 vs 496 voxels; +7.1%), whereas DCM patients showed reduction in activation with higher stimulation (1028 vs 1220 voxels; -2.2%). This activation pattern was consistent across the cervical spinal segments, with HCs peaking at C7 (sub-motor: 328 voxels, motor: 608 voxels), which is consistent with median nerve input for forearm and hand. While DCM responses were altered, with sub-motor differentially maximized at C6 (406 voxels) and abnormal recruitment at C5 (112 voxels) and C8 (198 voxels). Spatially, HC responses were focused in the gray matter (GM) and bilateral, whereas DCM responses showed reduced GM localization and were more ipsilateral. Furthermore, we saw spikes in sub-motor DCM deactivation at the overall (3032 voxels) and C6 root level (1500 voxels) that were greater than the other conditions. This study demonstrates the feasibility and effectiveness of median nerve stimulation as a task-based paradigm for assessing spinal cord function in those with spinal cord compression. Group level differences between DCM patients and age-matched HCs showed spatial and quantitatively distinct patterns across stimulation thresholds and spinal segments. Our findings suggest disrupted sub-motor and motor activation in the DCM group, highlighting the use of this approach as a functional biomarker to capture disease specific alterations.

Characterization of the hemodynamic response function in default mode network subregions for task-evoked negative BOLD responses
While the hemodynamic response function (HRF) of the positive blood oxygenated level dependent (BOLD) signal is well characterized in the functional magnetic resonance imaging (fMRI) field, studies investigating the shape and properties of the negative BOLD HRF, often observed in the default mode network (DMN) regions, are rare. In this study, we first investigated the linearity of the task-evoked negative BOLD response (NBR) with respect to stimulus duration. Then, we obtained the shape of the HRF from each region of the DMN, using an in-house developed unbiased and robust iterative deconvolution technique. Our results demonstrated that each region of the DMN represented a unique HRF, which was not only substantially different from the HRF of the positive BOLD signal, but also different from the HRF extracted from the other DMN regions. We replicated these findings using different fMRI datasets with distinct task paradigms. When comparing the HRF across DMN sub-regions and across tasks, our results demonstrated a significantly higher inter-regional variability compared to inter-task variability. Furthermore, our performance correlation analysis illustrated that the HRFs in the DMN sub-regions extracted by our iterative FIR method were not biased by the task performance level. Altogether, we demonstrated the linearity of the NBR, introduced a robust HRF extraction approach to obtain a distinct HRF for each DMN sub-regions, which were different from the HRF of the positive BOLD signal. The results suggest there is possibility that each node of the DMN might have different underlying neural and/or vascular mechanisms.

Comparing Neuroprotective Drug Efficacy in Rodent Neonatal Brain Injury Models
Background: A challenge in preclinical neonatal neuroprotection research is implementation of study designs that enable direct comparison of multiple potentially effective drugs. We used adaptive design to test four FDA approved drugs (azithromycin, erythropoietin, caffeine, melatonin) concurrently and determine which best combined safety and efficacy. Methods: Seven-day-old (P7) rats underwent hypoxia-ischemia (HI; right carotid ligation + timed 8% O2 exposure); some experiments included pre-treatment with agents that induced inflammation, and some included post-HI brief moderate hypothermia. Sensorimotor and neuropathology measures were incorporated into a Composite Score that also accounted for deaths. Outcome was initially evaluated at P21 and in confirmatory studies at P35. A pre-specified Bayesian algorithm with futility and efficacy stopping rules was devised to analyze emerging data and adjust subsequent animal allocation among drug groups. Results: In all models either azithromycin or erythropoietin (EPO) offered superior neuroprotection at P21 and the other was 'runner-up'. Caffeine and melatonin conferred modest neuroprotection in pure HI but were quickly eliminated in hypothermia-treated HI. At P35 azithromycin and EPO outcomes were generally similar. Conclusion: These results support azithromycin or erythropoietin as candidate neuroprotective agents and warrant future studies in large animal neonatal cerebral hypoxia-ischemia models.

Tentorium cerebelli notch severity in the human brain: protocol, distribution and histopathology validation
The dura forms an impression, the tentorial notch (TN), on the surface of the entorhinal cortex (EC). To characterize the potential injury to the EC, we evaluated the entorhinal surface in 55 postmortem samples. We developed a semi-quantitative protocol to rate the TN severity and evaluated the impact of the TN on the surrounding tissue. We rated the penumbral injury as: none, moderate or severe. We demonstrate that 96% of the samples showed a TN; specifically mild (33%), moderate (47%) and severe (16%). TN severity correlated positively with penumbra scores (Kendalls coefficient = 0.65, 95% CI [0.535, 0.735]). We found no association between TN severity levels and Braak and Braak stages (Fischers exact P value = .862). We established a robust protocol for assessing the tentorial notch, addressing a gap in understanding anatomical factors that may contribute to EC vulnerability in Alzheimers disease.

Effects of mind-wandering on cognitive and neural processes: Identifying specific impairments in adults with attention-deficit/hyperactivity disorder
Attention-deficit/hyperactivity disorder (ADHD) is a common neurodevelopmental disorder characterized by symptoms of inattention and/or impulsivity/hyperactivity. Although not a core symptom, excessive mind-wandering (MW) is commonly associated with the disorder and may contribute to symptoms as well as performance deficits observed in these patients. The main aim behind the present study is to better characterize MW and its impact on performance measures as well as on neural activity in adults with ADHD. Twenty-eight medication-naive ADHD patients and 28 healthy controls were matched on sex, age, and level of education; they completed a prolonged Go/No-Go sustained attention task including thought-probes to detect episodes of MW. Analysis of EEG data focused on P100, P3b, and CRN event-related potentials reflecting early visual, attention allocation, and performance monitoring processes, respectively. MW, particularly spontaneous MW, was more frequent in patients and linked to a more severe ADHD symptomatology. A differentiated negative impact of MW on performance was noted in both groups: speed and variability of patients' reaction times and accuracy in controls were affected. If the same perceptual (P100) and attentional (P3b) impairments were evidenced in both groups during MW, then the impairment of the performance monitoring (CRN) process was exclusively highlighted in ADHD patients (as revealed by a state x group interaction). Taken together, these results extend the understanding of the cognitive and neural processes associated with increased MW in patients and suggest performance monitoring as a potential key mechanism underlying specific MW-related impairments in ADHD.

Tissue Oxygenation Dynamics During Seizure to Spreading Depolarization in Rat Brain
Objective: Spreading depolarization is a phenomenon underlying various neurological conditions, including epilepsy. Researchers have suspected that local tissue oxygenation breakdown induces spontaneous SD. In this study, we investigated the relationship between spontaneous epileptic seizures and SD, with a focus on the role of local tissue oxygenation during the transition from seizure to SD. Methods: We applied a long pulse voltametric method to characterize local tissue oxygenation and extracellular space volume in the hippocampus of freely moving epileptic rats (male 6, female 3). Recordings were performed during the normal state of vigilance, spontaneous seizures, and seizure-associated SD events, as well as during their transitions. Results: No significant breakdown in local tissue oxygenation of HC was detected before SD onset during the seizure-to-SD transition. In contrast, a decreased ECS volume in the HC was observed before SD onset during this transition. Significance: Using a novel electrochemical approach in freely behaving rats with intact cerebral autoregulation, we demonstrate that ECS shrinkage, rather than breakdown of local tissue oxygenation, plays a leading role in SD initiation during seizure to SD transition. These findings refine our understanding of the mechanisms driving seizure-related SD and suggest that ECS dynamics may represent an important therapeutic target in epilepsy and other SD-associated neurological disorders.

Subretinal aspects of the optoretinographic response
Water movement in the living human retina and its regulation are important components of the tissue's structural integrity, optical properties, and homeostasis. In the outer retina there is a continuous flow of water from the subretinal space, through the retinal pigmented epithelium, and into the choroid. This flow is disrupted acutely in disorders such as retinal detachment and central serous retinopathy, and is also known to reduce dramatically with age and age-related macular degeneration. Optoretiongraphy is an emerging technique for measuring neural function in the retina by monitoring nanometer-scale deformations of the membranes of photoreceptors. These deformations have been hypothetically attributed, in part, to osmotic shifts that cause water to move into and out of the photoreceptor outer segment after light stimulation. In the present work, we describe a method for measuring changes in the lengths of the cone outer segment and subretinal space in parallel and results showing that light stimuli change the volume of the subretinal space. These results are consistent with earlier ex vivo measurements of its light-induced hydration. The magnitude of the latter changes depend on the rate of water clearance from the subretinal space, and thus may serve as an indicator of the health of the water transport system. In addition, they may help us understand the mechanisms underlying the photoreceptor optoretinogram. These findings add to a growing understanding of the ways in which light exposure leads to transient reconfigurations of the outer retinal layers lasting milliseconds to hours.

Brain Microstructure and Obesity Risk in Early Childhood: Insights from Restriction Spectrum Imaging
Pediatric obesity is a growing public health concern, yet little is known about the neurobiological underpinnings of obesity risk in early childhood. Using restriction spectrum imaging (RSI), we examined associations between adiposity and brain microstructure in a cross-sectional sample of 159 children aged 4-7 years, including 81 with ADHD and 78 typically developing (TD) peers. We focused on RSI-derived measures of restricted diffusion-specifically restricted normalized isotropic (RNI), directional (RND), and total (RNT) signals-as indicators of cellular density in subcortical and cortical regions implicated in reward and salience processing. Body mass index (BMI), percent body fat, waist circumference, and obesity status were assessed. Higher BMI, but not other adiposity measures, was significantly associated with increased RNI in the right insula, nucleus accumbens (NAcc), and putamen, as well as increased RNT in the right insula and pallidum. These findings suggest early microstructural alterations in reward-related circuits, consistent with theories of diet-related neuroinflammation. Contrary to hypotheses, ADHD diagnosis did not moderate the associations, and anthropometric profiles were similar between groups. This suggests a shared neural pathway linking early adiposity and brain structure, independent of diagnostic status. Our findings replicate and extend prior work in older children, highlighting BMI as the most sensitive marker of obesity-related brain differences in early childhood. These results underscore the potential of RSI as a tool for identifying early neural risk markers of obesity and inform future efforts to design preventive interventions during critical developmental windows.

Discovering flexible codes for prediction across timescales in the retina
Efficient coding theory postulates that a sensory system maximizes information between its response and the input, yet it is unclear if a different measure of optimality that takes into account output function might give a better fit to neural data. The sensory processing delays in many systems suggest that the maximization of predictive information is a reasonable objective function for driving fast, effective downstream behavior. We introduce a one-parameter family of optimal encoding distributions based on how far out in time a population of retinal ganglion cells is optimized to predict future stimuli. Analyzing the population response to a moving bar stimulus with rich temporal correlation structure identifies which particular optimal encoding best describes the neural activity. This allows for the discovery of how far out in time the retina is predicting, instead of simply testing for optimality at one timescale. As stimulus statistics change, so too does the time scale of prediction that best matches the population response. Focusing on this optimal timescale, the neural code can be evaluated in terms of classic efficient coding theory, revealing that the code also shows a peak in how these predictive bits are allocated in the population response repertoire. The stimulus has a fully controlled set of temporal statistics, but is still complex enough to show behaviors like starts and stops, constant motion, and motion reversals. Its tractable statistical structure allows for an information theoretic account of computations like motion anticipation and the retina reversal response in terms of the maximization of predictive information.

Long-acting naltrexone restores network connectivity in subjects with co-morbid cannabis and opioid use disorder
Co-morbid substance use disorders (SUDs) are common but difficult to study due to the complex, interacting, and overlapping mechanisms through which they affect brain networks. Many datasets collected to investigate a specific SUD include participants with co-morbid SUDs. While most studies treat comorbid SUDs as covariates of no interest, these covariates also contain untapped information. This is particularly relevant as cannabis use disorder (CanUD) has become increasingly prevalent and co-morbid with other SUDs that have been more thoroughly studied. While treatments have been established for multiple SUDs, none have been approved for CanUD, although naltrexone (NTX) has been associated with reduced use. Here, we conducted a retrospective secondary analysis of functional magnetic resonance imaging (fMRI) data from individuals with primary opioid use disorder (OUD) with co-morbid CanUD, alcohol use disorder (AUD), or cocaine use disorder (CocUD), while controlling for opioid use. All participants underwent imaging prior to receiving a therapeutic dose of long-acting intramuscular NTX (Vivitrol), an approved treatment for OUD and AUD but not for CocUD, and again two weeks post-administration. At baseline, OUD individuals with co-morbid CanUD, AUD, or CocUD exhibited distinct functional connectivity (FC) alterations compared to those with OUD-only. These differences were greater in younger participants and primarily involved the default mode network. Following NTX administration, FC differences between the co-morbid CanUD and OUD-only groups globally diminished. A similar FC response to NTX was observed in the parietal, subcortical, sensory, and cerebellar networks in the co-morbid AUD group. In contrast, little change in FC was observed in co-morbid CocUD. These findings, combined with prior evidence that NTX reduces cannabis use by dampening the experience of reward, suggest NTX may hold promise as a treatment for CanUD.

Cell Type-Specific Remodelling of the Female Hippocampus by Reproductive Experience and Age
Background: The hippocampus undergoes extensive cellular remodelling throughout life in response to multiple biological factors that shape its structure and function. Aging represents a fundamental driver of brain changes, with the hippocampus being particularly vulnerable to age-related concerns that contribute to cognitive decline and neurodegenerative disease risk. Reproductive experience, including pregnancy and motherhood, triggers profound hormonal fluctuations and metabolic demands that are increasingly recognized as major modulators of brain plasticity, yet represent an understudied and often paradoxical dimension of neurobiological variation. Although reproductive experience and aging are each known to influence biology across the dorsal and ventral hippocampus, their independent and interactive effects on cellular composition have not been systematically examined using quantitative approaches that can resolve cell-type-specific changes. Methods: We performed cell type deconvolution using Single-cell deconvolution of cell types (SCDC) on bulk RNA-sequencing data from female rat hippocampus, comparing nulliparous and parous females across young/older ages (7 month or 13 month old Sprague-Dawley rats, parous: 30 days or 6 months after giving birth) and dorsal/ventral regions. We harmonized 201 cell type annotations from three single-cell reference datasets into 23 biologically coherent categories utilizing female only data. Three-way ANOVA was used to identify independent and interactive effects. Complementary analyses (random forest, PCA, DESeq2) identified parity-associated transcriptional signatures. Cell-specific functional enrichment was performed by weighted gene set enrichment meta-analysis across multiple pathway databases and metrics. Results: Age emerged as the dominant factor affecting hippocampal cellular composition, significantly altering 6 harmonized cell types, particularly impacting microglia and oligodendrocytes. Regional effects were more extensive, affecting 9 harmonized cell types, while age x region interactions were minimal affecting two harmonized cell types. Critically, parity exerted independent effects on three cell populations: dorsal CA3 pyramidal neurons (p=0.0086-0.010 across two datasets) and SST interneurons (p=0.029) both showed 15-25% decreases, while astrocytes increased (p=0.022). Cell-type-specific pathway analysis revealed distinct molecular mechanisms with enrichment for protein degradation pathways in CA3 pyramidal neurons (GSK3B and BTRC:CUL1-mediated degradation; ubiquitin-dependent degradation of Cyclin D), neuroinflammation pathways in astrocytes, and GABAergic signaling disruption in SST interneurons (signaling receptor activity; GPCR ligand binding). Conclusions: Reproductive experience selectively remodels hippocampal cellular architecture through distinct, cell-type-specific molecular programs that operate independently of age and regional factors. These findings show parity as a critical biological variable requiring consideration in neuroscience and aging research and suggest specific cellular targets for understanding how reproductive history influences brain function.

Encoding neural representations of time-continuous stimulus-response transformations in the human brain with advanced deep neural networks
Human behavior arises from the continuous transformation of sensory input into goal-directed actions. While existing analytical methods often break time into discrete events, the stages and underlying representations involved in stimulus-response (S-R) transformations within time-continuous, complex environments remain only partially understood. Encoding models, combined with deep neural networks (DNNs) for feature generation, offer a promising framework for capturing these neural processes. While DNNs continue to improve in performance, it remains unclear whether these advances translate into more accurate models of brain activity. To address this, we collected fMRI data from participants (N = 23) as they played arcade-style video games and applied DNN-based encoding models to predict voxel-level brain activity. We compared the prediction accuracy of features from three DNNs at different stages of development within our encoding model. We show that the most advanced DNN provides the most predictive feature space for neural responses, while also exhibiting a closer hierarchical alignment between its internal representations and the brains functional organization. These results enable a more fine-grained characterization of time-continuous S-R transformations in high-dimensional visuomotor tasks, progressing along the dorsal visual stream and extending into motor-related regions. This approach highlights the potential of machine learning to advance cognitive neuroscience by enhancing the ecological validity of experimental tasks.

Low-frequency transcranial magnetic stimulation offers both immediate and delayed neuroprotection in neonatal hypoxia-ischemia model
Perinatal hypoxic-ischemic encephalopathy (HIE) is a leading cause of morbidity and mortality, and the current standard of care, therapeutic hypothermia, provides only partial neuroprotection. This study investigates the potential of low-frequency transcranial magnetic stimulation (LF-TMS) as a novel non-pharmacological adjunct therapy by targeting a key pathological mechanism of HIE: a persistent, pathological increase in glutamatergic synaptic transmission, or hypoxic long-term potentiation (hLTP). Using a neonatal mouse model of hypoxia-ischemia, we administered a single session of LF-TMS shortly after the hypoxic event. We then evaluated its effects on synaptic function via slice electrophysiology and on brain injury volume using serial MRI. Our results show that hypoxia-ischemia induced a significant and lasting synaptic potentiation in the brain's penumbral region. A single LF-TMS treatment successfully reduced this elevated glutamatergic response to control levels, suggesting a therapeutic mechanism similar to long-term depression (LTD) by regulating AMPA receptor redistribution. Furthermore, LF-TMS provided significant neuroprotection, as demonstrated by a reduction in the volume of the ischemic core and penumbra 48 hours after the injury. LF-TMS did not alter excitability in sham-treated mice, confirming its safety as a targeted intervention for pathological conditions without affecting normal brain function. This study provides strong evidence that LF-TMS is a promising neuroprotective strategy that acutely and subacutely mitigates brain injury in a neonatal hypoxia-ischemia model.

Frequency-Aware Interpretable Deep Learning Framework for Alzheimer's Disease Classification Using rs-fMRI
Gaining insight into the spectral and temporal alterations in brain connectivity associated with Alzheimer's disease (AD) may offer pathways toward more informative biomarkers and a deeper understanding of disease mechanisms. We propose FINE (Frequency-aware Interpretable Neural Encoder), a novel deep learning model designed to capture multi-scale temporal and frequency-specific patterns in dynamic functional network connectivity (dFNC) derived from resting-state fMRI. FINE integrates multiple expert branches, including convolutional layers, learnable wavelet layers, transformers, and static encoders, enabling the joint modeling of temporal evolution and spectral content of brain networks in an end-to-end framework. Beyond classification, FINE supports frequency-wise interpretability by aligning gradient-based saliency maps with statistical group differences, revealing potential robust, biologically meaningful biomarkers of AD. Evaluated on the large OASIS-3 dataset (856 subjects), FINE achieves AD classification performance (ROC-AUC 0.769) and provides insights into frequency-specific connectivity disruptions, particularly within subcortical, sensorimotor, and cerebellar networks. Our results demonstrate that incorporating frequency-aware modeling and interpretable architectures can advance both disease classification and underlying functional disruption of AD-related brain dynamics.

Explaining attractive and repulsive biases in the subjective visual vertical
Perception of gravity can be assessed by measuring the subjective visual vertical (SVV), the visually indicated spatial direction that appears earth-vertical to an observer. When the SVV is measured in darkness while the observer is roll-tilted, it shows substantial biases. At tilts larger than 45{degrees}, the bias is attractive, that is, the visual indicator appears vertical when rotated toward the observer. At smaller tilts, however, a repulsive bias is observed. The attractive bias has been explained within the Bayesian framework as the effect of a prior for upright posture. The repulsive bias has so far been considered anti-Bayesian, suboptimal, or as the result of uncompensated ocular counterroll. Here we show that both biases can be explained within a purely Bayesian model. More specifically, the repulsive bias at small roll-tilts is a consequence of the known tilt-dependent variability of the SVV, which is hypothesized to reflect different levels of sensory noise of the otolith organs. We thus provide a solution to a century-old question of why there is a repulsive bias in vertical perception.

Prediction of treatment response in infantile epileptic spasms syndrome using EEG phase-amplitude coupling
Objective: Treatment selection for infantile epileptic spasms syndrome (IESS) is complex and multifaceted, and currently no EEG biomarkers can guide this decision by predicting treatment response. We tested the predictive value of phase-amplitude coupling (PAC), as IESS patients are known to have elevated PAC. Methods: We analyzed retrospective EEG recordings from 40 IESS patients, before and after treatment, and 20 healthy controls. Patients were classified as responders (n=25) or non-responders (n=15) based on short-term treatment outcomes. We measured PAC in each EEG using modulation index (MI) and mean vector length (MVL) and analyzed the relationship between pre- and post-treatment values and the ability of pre-treatment values to predict response. Results: MI and MVL values decreased with treatment in almost all subjects. However, non-responders had significantly higher pre-treatment MI than responders (P<0.05), suggesting utility for predicting treatment response. Logistic regression modeling suggested that a 0.5 unit decrease in log(MI), which is approximately one IQR of the pre-treatment log(MI) values, results in a 6-fold increase in odds of positive treatment response. Significance: MI reflects short-term treatment response and is a candidate predictive EEG biomarker for IESS. MI may offer individualized insights for treatment selection and management strategies for IESS.

Local translation of circular RNAs is required for synaptic activity and memory
Circular RNAs (circRNAs) are enriched in synapses and implicated in cognitive processes, and recent studies have shown that circRNAs can encode micropeptides, which suggests that there may be novel synaptic proteoforms in the brain that await discovery. Here we report widespread learning-induced local circRNA translation in the prefrontal cortex of male C57BL/6 mice. More than 1500 synapse-enriched circRNAs contain active IRES elements, with 842 interacting with the ribosome and 241 exhibiting direct evidence of activity-induced translation. We discovered a synapse-enriched micropeptide (P1) that is derived from a single exon circRNA, the mRNA host of which encodes an enzyme associated with protein repair. Although P1 is only a third of the size of the full-length protein, it is locally expressed, enzymatically active, and interacts with plasticity-related proteins, including CaMKII. In addition to direct effects on synaptic activity, targeted P1 knockdown impairs whereas its overexpression enhances fear extinction memory. These findings shed new light on the "dark" proteome in the brain and reveal local circRNA translation as a novel mechanism of plasticity and memory.

A mobile circular DNA element drives memory-related processes in mice
Extrachromosomal circular DNAs (ecDNAs) are double stranded, closed loop DNA molecules that act independently of the genome. Here we report the discovery of a brain-enriched ecDNA, ecCldn34d, that regulates the transcriptional state of plasticity-related genes and is temporally associated with the stability of fear memories. ecDNA therefore represents a novel class of mobile DNA elements that are involved in experience-dependent neuroadaptation.

Mobile Eye-Tracking Glasses Capture Ocular and Head Markers of Listening Effort
To extend the assessment of listening effort beyond a sound booth, we validated mobile eye-tracking glasses (Pupil Labs Neon) by comparing them to a stationary system (Eyelink DUO) in a controlled environment. We recorded eye movements, pupil size, and head movements from 26 young adults during a speech-in-noise task. When listening conditions became challenging, we observed reduced gaze dispersion and increased pupil sizes of similar magnitude from both devices, in addition to reduced head movements recorded solely by the mobile device. These findings suggest that mobile eye-trackers reliably capture listening effort, paving the path towards assessments in daily settings.

Pre-stimulus Brain States Predict and Control Variability in Stimulation Responses
Does the ongoing brain state determine how it responds to localized electrical stimulation? This fundamental question has major implications for neuroscience and medicine, as stimulation outcomes remain highly variable even when identical parameters ("how") and target sites ("where") are used. Such unpredictability undermines reproducibility, limits clinical reliability, and forces current protocols to rely on empirical trial-and-error rather than principled, evidence-based strategies. Mounting evidence suggests that the brain's state prior to stimulation--a crucial "when" factor--shapes responses, yet the most reliable predictive markers remain unknown. Here, we systematically characterize the links between pre-stimulus (spontaneous) activity and post-stimulus (evoked) responses using simultaneous high-density EEG and stereotactic EEG from 36 epilepsy patients across 379 sessions (21,800 stimulations). We show that large-scale neural dynamics robustly predict stimulation outcomes, with a subset of measures--particularly network synchronization, functional connectivity, and spatiotemporal signal diversity--consistently forecasting responses across sessions. Whole-brain activity enhanced prediction compared to local assessments, and predictability varied across networks, being strongest in sensorimotor and visual regions. These findings establish a quantitative framework for state-dependent brain stimulation: by timing interventions to optimal pre-stimulus states, variability can be reduced and reproducibility enhanced. Our results directly address the fundamental question of where and when to stimulate, providing a pathway toward evidence-based protocols with improved therapeutic precision.

Psychedelics Relax Priors and Reshape Orbitofrontal Dynamics
Psychedelics such as psilocybin and ketamine are gaining attention as rapid-acting treatments for psychiatric disorders, yet the mechanisms by which they alter cognition remain unclear. A key hypothesis from the REBUS model proposes that psychedelics relax high-level priors, allowing bottom-up sensory information to exert greater influence over perception and behavior. Here, we test this model in mice performing a free-response perceptual decision-making task that disambiguates prior-driven and sensory-driven decision strategies. Acute administration of psilocybin or ketamine significantly slowed decision times and improved accuracy. Behavioral modeling that combined drift diffusion and GLM-HMM frameworks revealed that these changes were mediated by increased decision thresholds and a marked shift into sensory-engaged cognitive states. Whole-brain c-Fos mapping identified a distributed decision-making network, with psychedelics selectively modulating cortical and subcortical nodes. Calcium imaging in the orbitofrontal cortex (OFC), a key region for integrating priors and sensory inputs, revealed preserved decision-related selectivity under psychedelics, while exhibiting reduced neuronal correlations, i.e. population-level signatures of weakened top-down influence and relaxed priors. Together, these results provide circuit-level support for the REBUS model, showing that psychedelics reconfigure brain-wide and local dynamics to promote more deliberate, flexible, and sensory-driven decision policies.

Cereblon-related mild intellectual disability disrupts response inhibition and uniformity of group--individual strategies
Temporal processing, including duration, is essential for survival and communication across species. Intellectual disability (ID), which has diverse causes, including Cereblon (CRBN), impairs duration discrimination. CRBN-related ID, link to abnormal cognitive behaviors, may disrupt both perception and behavior during duration discrimination. However, cross-species behavioral strategies, their variation with ID, and associated behavioral indices remain unclear. Here, humans and wild-type (WT) mice with typical intelligence, and CRBN knockout (KO) mice with ID, performed an auditory duration discrimination task with Long (10 s) and Short (2 s) cues. All groups distinguished stimulus durations, but latency-based strategies diverged. KO mice showed impulsivity and divergent responses at both individual and group levels, whereas WT mice and humans consistently delayed their responses by ~2 s to the Short cue length, reflecting inhibition and convergent responses. In typical intelligence models, latencies for both stimuli clustered between 2--5 s, while in ID model depended on stimulus duration. Duration perception is a conserved cross-species capacity, while task-specific cognitive strategies are intrinsically preserved though variation emerges with ID. We suggest that latency indices dissociated inhibition from impulsivity, and convergence from divergence. We further suggest that CRBN-related ID preserves perceptual understanding but disrupts the shared behavioral language of typical intelligence.

Genomic plasticity drives olfactory adaptation in a pest fly
Preference shifts in insects are often driven by changes in the olfactory system, yet the underlying mechanisms remain unclear. The worldwide pest Drosophila suzukii, which oviposits in ripe rather than overripe fruits, provides a powerful model to study these mechanisms and their behavioral consequences. Here, we show that this shift is linked to functional remodeling in four olfactory receptor neurons (ORNs): ab2B, ab3A, ab4B, and ab10A. While ab3A and ab10A exhibit tuning changes shared with the non-pest relative D. biarmipes, ab2B and ab4B display species-specific adaptations in D. suzukii. These changes result not only from receptor sequence divergence but also from novel innovations: receptor co-expression in ab3A and partitioned expression of Or67a paralogs in ab2B and ab10A. Together, these findings show how genomic plasticity in chemosensory gene families enables rapid sensory adaptation and niche transition.

Cancer-induced Nerve Injury Unveils a Sympathetic-to-Sensory Nerve Axis in Head and Neck Cancer
Oral squamous cell carcinoma (OSCC) is one of the most painful cancers, with patients frequently reporting spontaneous, neuropathic-like pain. While sympathetic and sensory nerves have been individually implicated in cancer progression, whether and how these systems interact to drive pain and tumor growth has remained unclear. Here, we integrate prospective human data with reverse-translational mouse models to reveal that cancer-induced nerve injury unveils crosstalk between sympathetic postganglionic neurons and trigeminal sensory afferents in the tumor microenvironment. In patients, circulating norepinephrine (NE) correlated with spontaneous pain and perineural invasion, identifying a potential sympathetic contribution to disease burden. In mice, aggressive non-immunogenic OSCC tumors evoked spontaneous nociceptive behaviors, elevated tumoral NE, and sensory nerve injury marked by ATF3 expression and hyperexcitability. Tumor-associated sensory neurons acquired adrenergic sensitivity through 1-adrenergic receptor plasticity, while sympathetic neurons exhibited plasticity characterized by sprouting, altered gene expression, and heightened excitability, creating a maladaptive feed-forward loop that amplified nociceptive signaling. Disrupting this sympathetic-sensory communication by sympathectomy or selective ablation of TRPV1 sensory fibers reduced tumor growth, sympathetic tone, and spontaneous pain like behaviors, although sensory adrenergic sensitivity persisted. Together, these findings establish that reciprocal sympathetic-sensory plasticity and crosstalk in the tumor may fuel both OSCC progression and neuropathic like pain. Targeting this peripheral neuroplasticity may offer a translational strategy to limit tumor growth and alleviate pain.

Vascular-Perfusable Human 3D Brain-on-Chip
Development and delivery of treatments for neurological diseases are limited by the tight and selective human blood-brain barrier (BBB). Although animal models have been important research and preclinical tools, the rodent BBB exhibits species differences and fails to capture the complexity of human genetics. Microphysiological systems incorporating human-derived cells hold great potential for modeling disease and therapeutic development, with advantages in screening throughput, real-time monitoring, and tunable genetic backgrounds when combined with induced pluripotent stem cell (iPSC) technology. Existing 3D BBB-on-chip systems have incorporated iPSC-derived endothelial cells but not the other major brain cell types from iPSCs, each of which contributes to brain physiology and disease. Here we developed a 3D Brain-Chip system incorporating endothelial cells, pericytes, astrocytes, neurons, microglia, and oligodendroglia from iPSCs. To enable this multicellular 3D co-culture in-chip, we designed a GelChip microfluidic platform using a 3D printing-based approach and dextran-based engineered hydrogel. Leveraging this platform, we co-cultured and characterized iPSC-derived brain-on-chips and modeled the brain microvasculature of APOE4, the strongest known genetic risk factor for sporadic Alzheimers disease. These 3D brain-on-chips provide a versatile system to assess BBB vascular morphology and function, investigate downstream neurological effects in disease, and screen therapeutics to optimize delivery to the brain.

Multisensory attenuation of the pupil light response in autistic and non-autistic children
Autonomic responses to sensory stimuli are altered in autism, yet little is known about how multisensory input modulates these responses. This study examined whether auditory stimuli affect the pupil light reflex (PLR), a parasympathetically driven response to light, in autistic and non-autistic children. Pupillometry was used to measure responses to visual-only (V), auditory-only (A), and audiovisual (AV) stimuli in 72 children aged 6-14 years (34 non-autistic, 38 autistic). We hypothesized that auditory input would attenuate pupil constriction in non-autistic children and that this cross-modal modulation might differ in autism, reflecting altered sensory-autonomic functioning. Across groups, results revealed a consistent pattern: auditory stimuli elicited pupil dilation, visual stimuli evoked constriction, and simultaneous audiovisual stimuli led to attenuated constriction relative to visual-only trials. This attenuation lends support to prior findings of multisensory attenuation of the PLR. Time-binned analysis revealed a group effect during the 500-1000 ms post-stimulus window: autistic children showed significantly more positive baseline-corrected pupil responses across conditions (i.e., less constriction in V/AV and greater dilation in A), suggesting group differences in the dynamic trajectory of the pupil response. Contrary to expectations, autistic and non-autistic children did not differ significantly on peak constriction or constriction latency within visual conditions. Findings support the presence of cross-modal modulation of the PLR in both autistic and non-autistic children and suggest that auditory signals influence early-stage visual-autonomic processing similarly across groups. Pupillometry may provide a promising, noninvasive tool for probing sensory-autonomic interactions in autism. Future studies with paradigms optimized for pupil measurement may reveal more nuanced group differences and clarify links to real-world sensory challenges.

Optimizing and assessing multichannel TMS focality
Background: Multichannel transcranial magnetic stimulation (mTMS) enables electronic steering of induced electric fields across multiple cortical targets without physical coil repositioning, addressing key limitations of conventional single-channel TMS (sTMS). However, determining optimal input currents for focal stimulation remains challenging, and different mTMS systems have not been systematically compared under realistic hardware constraints. Objective: To develop an user-centric framework for optimizing and assessing mTMS focality by introducing a generic optimization algorithm, establishing meaningful focality metrics, and comparing mTMS coil arrays with traditional single-channel TMS across cortical targets. Methods: We developed a fast optimization framework incorporating target E-field constraints via parametrization of degenerated hyperellipsoids, explicitly integrating current-rate limits, for example from stimulator electronics and coil heating. Using high-resolution finite-element models of nine individual brains, we compared two mTMS designs (5-channel planar and 6/12-channel spherical systems) with standard sTMS figure-of-eight coils. Three complementary metrics quantified performance: Focality, Target2Max, and OverstimulatedArea. Results: Despite using a single optimized placement for all region-of-interest targets, mTMS achieved focality comparable to repositioned single-channel TMS. For superficial targets, single-channel TMS showed slightly better focality, but for deeper cortical targets (>25mm skin-cortex distance), mTMS performed similarly. More stimulation channels improved focality but required stronger current-rate constraints. The planar design performed better for deeper targets, while spherical designs improved with additional channels. Conclusion: mTMS systems demonstrate remarkable performance comparable to standard TMS, enabling efficient multi-target stimulation without repositioning. Our open-source framework provides practical tools for designing and evaluating mTMS systems, supporting goal-directed mTMS development and effective application.

Bodily Self in Macaque Monkeys and Human Infants
Whether monkeys possess the concept of bodily self remains controversial. We found that macaque monkeys trained for mirror self-recognition (MSR) could spontaneously recognize themselves when viewing video images of their back from a third-person perspective (3PP), but this generalization required a few minutes of "exploration". In contrast, through visual-tactile synchronization training similar to that used in human out-of-body studies, naive monkeys could recognize themselves from 3PP back images and demonstrated MSR within seconds. Interestingly, in human infants, we found that 3PP back self-recognition developed spontaneously several months after MSR. These results demonstrate the existence of hierarchical bodily self in macaques and similar stepwise development of bodily self in human infants, offering insights into the emergence and manifestation of bodily self in natural and artificial intelligence.

Secondary nucleation of α-Synuclein drives Mitochondria dysfunctions and Lewy body formation in Parkinson's Disease
The seeding of -Synuclein (Syn) is a key driver of Lewy pathology propagation in Parkinson's disease (PD) and forms the basis for recent diagnostic advances. However, it remains unclear how the structural and biochemical features of Syn seeds dictate their propagation efficiency, capacity to induce Lewy body formation, and resulting cellular toxicity. Using genetic and idiopathic PD cell models, we map the pathogenic cascade beginning with the seed-driven conversion of endogenous Syn, followed by impaired degradation, mitochondrial dysfunction, and ultimately Lewy body formation. By coupling kinetic modelling of aggregation with functional readouts, we identify secondary nucleation as the predominant mechanism generating toxic Syn aggregation intermediates, identifying the critical process that links seeding to pathology. Extending this framework to PD brain, we quantitatively correlate seeding capacity with the spatiotemporal spread and severity of Lewy pathology, revealing a mechanistic connection between Syn aggregation dynamics and disease progression at molecular, cellular, and anatomical levels. By unifying molecular mechanism with clinicopathological progression, our work identifies catalytic Syn fibrillar seeds as tractable targets for both disease-modifying therapy and biomarker development in PD.

Accelerated rTMS for Enhancing Intact Cognition: An Examination of Dose Effects on Electrocortical Indicators of Attention and Working Memory
BACKGROUND AND AIMS Improvements in cognition (e.g., attention, working memory) are common after repetitive trans-cranial magnetic stimulation (rTMS) treatment and have also been observed in non-clinical samples. This study investigated whether rTMS can enhance cognitive resilience in individuals who perform in high-stress environments using cognitive measures and electroencephalography (EEG) to explore potential neural mechanisms of rTMS-induced cognitive change. METHODS 40 college-educated adults not reporting cognitive or psychiatric concerns underwent a 5-day accelerated (10 sessions/ day) rTMS treatment and pre-post cognitive assessment. Participants were assigned to 1 of 10 doses defined as number of active vs. sham stimulation sessions/ day (total active pulses = 3,000-30,000). To assess cognitive effects, standard batteries (NIH Toolbox, Spaceflight Cognitive Assessment Tool for Windows [WinSCAT]) were administered-and a subset (n=21) also did an N-Back working memory task with EEG measurement-before, immediately after, and 1 month after rTMS. For all indices, linear and quadratic correlations of pre-to-post-rTMS change with dose were examined to test if an optimal dose was achieved. RESULTS From pre- to post-rTMS, participants improved in fluid cognition (working memory, processing speed) as measured by NIH toolbox, t(39)=8.4, p<.001, d=1.3, and WinSCAT, t(39)=4.1, p<.001, d=.65; and, improvement in the latter related linearly to rTMS dose, r(39)=.40, p=.01. Regarding EEG, subjects who received high (6+ active sessions/day) also showed increased amplitudes of an event-related marker of stimulus-directed attention (P300) whether it was elicited by simple task targets, t(10)=3.1, p=.01, d=.95, or task-unrelated noise stimuli played during the task as probes of peripheral attention, t(10)=3.6, p=.005, d=1.1. Additionally, there were linear relationships between change magnitude and rTMS dose for simple task target-related, r(20)=.45, p=.04, and peripheral noise-related, r(20)=.54, p=.009, P300s. CONCLUSIONS rTMS improved fluid reasoning abilities and also changed how dynamic attention is deployed during high-demand challenges as a potential mediator of fluid cognition improvements. While linear dose-response relationships support that changes were rTMS-elicited, absence of a response curve asymptote also suggests that still-higher doses could be warranted to achieve maximal effects in non-clinical samples.

Gait adaptation to asymmetric foot-ground compliance applied by robotic footwear
Perturbing foot-ground interaction dynamics has shown promise for eliciting adaptations to inter-limb weight-bearing symmetry, a critical target for rehabilitation of asymmetrical neuromotor deficits affecting gait. To date, this perturbation paradigm has been delivered via adjustable stiffness treadmills, in which the treadmill deck displaces under the foot during the stance phase of gait. Recently, we developed robotic footwear capable of delivering asymmetrical ground stiffness perturbations during walking, offering a portable experimental platform for asymmetrical ground stiffness perturbations, allowing it to be executed overground and on conventional treadmills. In this study, we quantified kinetic and spatio-temporal gait adaptation to a foot-ground compliance perturbation delivered by the novel robotic footwear. We found that participants shifted their vertical ground reaction force impulse, vertical pushoff peak, and peak braking force toward the unperturbed limb after the perturbation was removed, indicating a shift in neuromotor control elicited by the perturbation. We conclude that the robotic footwear can elicit weight bearing adaptations similar to adjustable stiffness treadmills, and that the dissipative mechanical properties of the shoes likely play a key role in the direction of the adaptation.

Spatial predictive coding in visual cortical neurons
Predictive coding is a theoretical framework that can explain how animals build internal models of their sensory environments by predicting sensory inputs. Predictive coding may capture either spatial or temporal relationships between sensory objects. While the original theory by Rao and Ballard, 1999 described spatial predictive coding, much of the recent experimental data has been interpreted as evidence for temporal predictive coding. Here we directly tested whether the "mismatch" neural responses in sensory cortex are due to a spatial or a temporal internal model. We adopted two common paradigms to study predictive coding: one based on virtual-reality and one based on static images. After training mice with repeated visual stimulation for several days, we performed multiple manipulations, including: 1) we introduced a novel stimulus, 2) we replaced a stimulus with a novel gray wall, 3) we duplicated a trained stimulus, or 4) we altered the order of the stimuli. The first two manipulations induced a substantial mismatch response in neural populations of up to 20,000 neurons recorded across primary and higher-order visual cortex, while the third and fourth ones did not. Thus, a mismatch response only occurred if a new spatial -- not temporal -- pattern was introduced.

Sensitivity of the human temporal voice areas to nonhuman primate vocalizations
In recent years, research on voice processing in the human brain - particularly the study of temporal voice areas (TVA) - was dedicated almost exclusively to conspecific vocalizations. To characterize commonalities and differences regarding primate vocalization representations in the human brain, the inclusion of closely related nonhuman primates - namely chimpanzees and bonobos - is needed. We hypothesized that neural commonalities would depend on both phylogenetic and acoustic proximities, with chimpanzees ranking closest to Homo. Presenting human participants (N=23) with the vocalizations of four primate species (rhesus macaques, chimpanzees, bonobos and humans) and regressing-out relevant acoustic parameters using three distinct analyses, we observed within-TVA, sample-specific, bilateral anterior superior temporal gyrus activity for chimpanzee vocalizations compared to: all other species; nonhuman primates; human vocalizations. Within-TVA activity was also observed for macaque vocalizations. Our results provide evidence for subregions of the TVA that respond principally - but not exclusively - to phylogenetically and acoustically close nonhuman primate vocalizations, namely those of chimpanzees.

Population coupling of V1 and V4 neurons and its relation to local cortical state fluctuations and attention in macaque monkey
Neurons couple to various degrees to the activity level of the local neighboring population whereby strongly coupled 'choristers' and weakly coupled 'soloists' have been identified as two extremes of a continuous spectrum. At the same time neuronal populations undergo coordinated On and Off cortical state activity fluctuations, which are locally modulated by attention. The population coupling of soloists and choristers suggests that soloists should show limited alignment with cortical state fluctuations, while choristers should exhibit profound alignment. To test this, we recorded neurons across cortical layers in macaque areas V1 and V4, while animals performed a feature based spatial attention task. As expected, we found a wide range of population coupling strength of neurons. In line with our prediction, coupling of choristers to cortical state changes (ON-OFF transitions) was generally stronger than that of soloists. The strength of population coupling of neurons was similar during spontaneous and stimulus driven activity. Allocation of attention to the receptive field reduced the population coupling strength. Attentional modulation of neurons was positively correlated with population coupling strength. While neurons on averaged retained their coupling strengths across conditions, some neurons can change coupling strength condition dependent, thereby potentially enhancing the coding abilities of cortical circuits.

Verbal Episodic Processing in Newborns
During the first period of life, human infants rapidly and effortlessly acquire the languages they are exposed to. Although memory is central to this process, the nature of early verbal memory systems and the factors that determine retention and forgetting remain largely unknown. Behavioural and brain measures have demonstrated memory formation in newborns. However, word traces fade in the face of acoustic overlap, leading to interference and forgetting. Here, we investigate whether speakers' identity changes facilitate the separation into distinct acoustic episodes and the creation of non-overlapping verbal memories. Newborns (0-4 days-old) were tested in a familiarization-interference-test protocol, while neural cortical activity was recorded using functional Near-Infrared Spectroscopy (fNIRS). The results showed higher neural activation for novel words compared to familiar ones in the test phase, indicating that the infants recognized the familiar words despite the presence of potentially interfering sounds. The recognition response was measured over the left inferior frontal (IFG) and superior temporal gyrus (STG) areas, known to be crucial for encoding auditory information and language processing. The neural response also involved the right IFG and STG, involved in interpreting vocal social cues and speaker recognition. These data show that speaker identity is a key feature of speech, enabling episodic-like memories from birth and evolutionary advantages at the outset of human communication.

Theory of Mind and Discourse Production in Schizotypy: An fMRI Study
Background: Schizotypy (ST) reflects subclinical traits linked to schizophrenia spectrum disorders and associated cognitive and social impairments. Theory of Mind (ToM) and discourse production deficits are well documented in schizophrenia (SZ), yet the neural basis of discourse-related ToM processes in ST remains unclear. This study investigated brain activation during narrative planning and production in individuals with schizotypal traits. Methods: Thirty young adults (mean age = 18.8 years) completed standardized assessments, including the Schizotypal Personality Questionnaire Brief Revised (SPQBR), adverse childhood experiences (ACEs), depression (PHQ9), and dissociation (DESB). Participants performed a discourse task in an fMRI scanner, describing nine-panel cartoons requiring inference of character intentions. Behavioral discourse metrics included total and inferred events. fMRI analyses examined activation during planning and production phases, with SPQBR positive, negative, and disorganized traits entered as regressors. Results: Schizotypal traits correlated with multiple psychosocial risk factors, including elevated depression, ACEs, and dissociation (r = .48 .82, p < .01). During planning, canonical ToM/self-referential regions (vmPFC, precuneus, insula) were recruited. Positive traits correlated with increased activation in the right temporo-parietal junction, precuneus, and lingual gyrus, whereas disorganized traits were associated with reduced activation in the precuneus and lingual gyri. During production, networks spanning vmPFC, hippocampus, right TPJ, and basal ganglia were engaged. Negative traits correlated with increased motor/premotor activation, while disorganized traits correlated with reduced activation in lingual gyrus, SMA, and cerebellum. Conclusions: Findings demonstrate distinct neural correlates of schizotypal traits during discourse planning and production, supporting models of schizophrenia spectrum risk emphasizing disrupted inference and integration processes.

Memory-specific E-I balance supports diverse replay and mitigates catastrophic forgetting
The hippocampus is the brain's central locus for memory processing. In a widely accepted hypothesis, the hippocampus stores short-term memories in attractor states, which are then consolidated in the neocortex as long-term memories. Hippocampal replay activities play a pivotal role in memory consolidation, but memories are only transiently, not persistently, replayed, casting doubt on the attractor memory hypothesis. Concepts like memory capacity, the upper bound on the number of storable stable memories in a neural network, may also need to be revised if their offline transient activation is more essential for memory consolidation. Here, we explore the biologically plausible mechanism of offline memory processing by asking a recurrent spiking network to replay reliably as many stored assemblies as possible. We demonstrate the importance of memory-assembly-specific inhibitory plasticity for replaying diverse memory assemblies. Furthermore, our results highlight the role of transient memory states in mitigating catastrophic forgetting.

Cortical tension as a mechanical barrier to safeguard against premature differentiation during neurogenesis
Neuronal differentiation requires coordinated gene reprogramming and morphodynamic remodeling. How mechanical forces integrate with nuclear gene programs during neurogenesis remains unresolved. Here, we identify cortical tension as a mechanical barrier that safeguards against premature neuronal differentiation. Deletion of Plexin-B2, a guidance receptor controlling actomyosin contractility, lowers this barrier, enabling neurite outgrowth and accelerating neuronal lineage commitment. We show that coupling of extrinsic differentiation cues with intrinsic morphodynamics is essential for stabilizing neuronal fate and that cortical barrier and epigenetic barrier act in concert to regulate developmental timing. In cerebral organoids, Plexin-B2 ablation triggered premature cell-cycle exit and differentiation, resulting in progenitor pool depletion and neuroepithelial disorganization, phenotypes echoing intellectual disability in patients with rare pathogenic PLXNB2 variants. Our studies demonstrate that cortical tension functions as mechano-checkpoint that regulates the onset of neurogenesis. Lowering this barrier may provide a strategy to accelerate induced neuron generation and maturation for CNS disease modeling.

Network dynamics underlying activity-timescale differences between cortical regions
Network-level dynamics are thought to be central to computation in the cerebral cortex. Yet, how these dynamics differ across areas remains poorly understood. We leveraged an intrinsic property of cortical regions to tackle this problem - the timescales over which they spontaneously sustain activity. We first co-registered functional and spatial transcriptomics datasets to show that timescales across the mouse cortex are predicted by many transcript categories, including those that regulate circuit wiring. Next, we used simultaneous two-photon imaging and optogenetics in mice to ask how these putative differences in connectivity lead to distinct network responses to brief, focal excitatory input to a short-timescale visual area, VISp, and a long-timescale frontal area, MOs. MOs neurons were more likely to respond to photostimulation of their neighbors. Moreover, the evoked dynamics of the overall network were much longer lasting in MOs than VISp, due to the more prevalent recruitment of late-responding neurons, which formed reliable activity sequences. Overall, our findings show that, beyond single-neuron timescales, different cortical areas are distinctly wired to sustain input over varying time windows via network dynamics, with important implications for our understanding of cortical computation.

Neuroelectrophysiological correlates of extended cessation of consciousness in advanced meditators: A multimodal EEG and MEG study
In some contemplative traditions, "extended cessation" (EC) refers to a state of advanced meditation in which the meditator intentionally suppresses their own consciousness and re-emerges with a profound sense of clarity and equanimity. Here, we present the first electrophysiological study of EC, in which five meditators underwent concurrent electroencephalography (EEG) and magnetoencephalography (MEG) recording. EC significantly reduced alpha power. EC also tended to increase neural complexity, unlike other states in which consciousness is absent (e.g., sleep, anesthesia, disorders of consciousness). Our results indicate that the neural correlates of EC are distinct from other states that induce loss of consciousness and that complexity is not a sufficient condition for consciousness, while also providing new insights into the implications of advanced meditation for human flourishing.

The curriculum effect in visual learning: the role of readout dimensionality
Generalization of visual perceptual learning (VPL) to unseen conditions varies across tasks. Previous work suggests that training curriculum may be integral to generalization, yet a theoretical explanation is lacking. We propose an explanatory theory of visual learning generalization and curriculum effects by leveraging an artificial neural network (ANN) model of VPL in comparison with humans. We found that easy-to-hard sequential training improved generalization in both humans and ANNs. However, when easy and hard conditions were interleaved, humans and ANNs showed different behaviors: while ANNs performed worse than with sequential training, humans maintained good performance but with large inter-individual variability. Investigating ANN models trained with different curricula, we demonstrated that models relying on low-dimensional neural populations showed superior generalization. This readout subspace dimensionality was directly determined by curriculum: learners who learned from easy tasks early formed lower-dimensional subspaces and generalized better. Our theory provides a mechanistic framework linking curriculum design to VPL generalization through neural population dimensionality.

Eye-Head Coordination During Rapid Gaze Shifts in Soccer Scanning
The visual exploratory behavior involving rapid gaze shifts that supports cognitive processes is called scanning. Scanning in soccer is a foundational behavior that enables players to explore their environment, supporting rapid and accurate decision-making. However, since most previous studies have focused on head movements, the coordination structure of gaze shifts, including the ocular contribution, is insufficiently understood. This study aimed to determine the coordinated roles of the eye and head during scanning. Twenty male collegiate soccer players performed a passing task paired with video-based situational judgments while eye and head movements were recorded by a 200 Hz sampling rate using an eye tracker system with a built-in gyroscope. We detected eye and head velocities and amplitudes during rapid gaze shifts. As a result, peak eye velocity was significantly higher than peak head velocity (d = 4.09, p < .001). The cross-correlation (CC) between gaze and eye velocities (0.98) was significantly greater than that between gaze and head velocities (0.90). These results indicate that gaze shifts were driven primarily by eye movements rather than head movements. Furthermore, eye and head velocities were negatively correlated (r = -.60, p = .005), whereas their velocity-control profiles (mean velocity normalized by amplitude) were positively correlated (r = .90, p < .001), indicating individual strategy-dependent allocation of effort while preserving control capability of eye and head movements. Our findings provide a characterization of the eye-head coordination structure of gaze shifts in soccer scanning, suggesting the need for comprehensive assessments beyond head-motion metrics alone.

Tau oligomers modulate synapse fate by eliciting progressive bipartite synapse dysregulation and synapse loss
Background Synapse function is critical for cognition, and synapse loss is highly correlated with cognitive decline in Alzheimer's disease and related dementias. Tau oligomers, which accumulate in the brain in Alzheimer's disease, can acutely inhibit synaptic plasticity and cause synapse loss. Coordinated presynaptic and postsynaptic function is essential for effective synaptic transmission, and both compartments can be dysregulated by pathogenic tau. However, the series of pathophysiological events triggered by tau oligomers to cause the dysfunction and deterioration of presynaptic terminals and postsynaptic sites remain unclear. Methods We developed a proximity labeling tool to map the postsynaptic proteome by fusing PSD-95 with APEX2 (APEX2-PSD-95) which was expressed in human induced pluripotent stem cell (iPSC)-derived neurons. We used APEX2-PSD-95 to map the dynamic changes in the postsynaptic proteome with precise temporal resolution after an acute exposure of human iPSC-derived neurons to recombinant tau oligomers for 30 min. Leveraging immunocytochemistry, electrophysiology and electron microscopy, we further delineated the impact of the acute tau oligomer exposure on presynaptic and postsynaptic compartments over time for up to 14 days. Results The brief exposure of human iPSC-derived neurons to tau oligomers caused a progressive deterioration of synapses, marked by both presynaptic and postsynaptic dysregulation. Postsynaptic proteome mapping revealed an immediate tau oligomer-triggered downregulation of the postsynaptic actin motor proteins Myosin-Va and Myosin-10, which coincided with impaired AMPA receptor (AMPAR) trafficking during synaptic plasticity. This was followed 24 hours later by the upregulation of disease-related proteins, including GSK3B; at postsynaptic sites. The loss of PSD-95-labeled postsynaptic sites at 7 days after tau oligomer exposure preceded the loss of Synapsin-labeled presynaptic terminals at 14 days. The postsynaptic sites that remained exhibited a long-term downregulation of postsynaptic AMPAR levels and sustained synaptic plasticity impairment. Moreover, the remaining presynaptic terminals contained less clusters of vesicles at the presynaptic active zone which was associated with reduced vesicle release probability at synapses. Conclusion Our findings reveal the series of events underlying tau oligomer-induced bipartite synapse deterioration. The progressive decline of synapses involves the emergence of two synapse fates. One synapse fate involves the persistent weakening of both presynaptic and postsynaptic function, and the other results in synaptic loss.

The use of Artificial Intelligence in Magnetic Resonance Imaging of Epilepsy: A Systematic Review and Meta-Analysis
Background. The application of artificial intelligence (AI)/machine learning (ML) to MRI can be a powerful tool to streamline clinical decision-making, yet variability amongst MRI sequences and algorithms have hindered appropriate assessment of reliability and generalizability. Methods. We conducted a systematic review and meta-analysis of the ability of current AI/ML models operating on MRI data for: 1) epilepsy diagnosis, 2) temporal lobe epilepsy lateralization, 3) lesion localization, and 4) post-surgical outcome prediction. Searches were conducted across PubMed, Medline, and Embase databases from inception until January 1, 2025. We selected studies that employed AI/ML models trained on any MRI modality to classify at least ten patients across the four main objectives for qualitative assessment, and further included in the meta-analysis if they reported an accuracy rate. The primary outcome of the meta-analysis was the overall accuracy of AI/ML models trained on MRI data. The secondary outcome was the concomitant risk of bias evaluation using PROBAST. Results. We identified 158 studies for qualitative evaluation and 127 studies for inclusion in the meta-analysis. AI/ML on multimodal MRI could accurately distinguish epilepsy patients from healthy controls (overall accuracy: 88% [85-90]), lateralize temporal lobe epilepsy (90% [87-93]), localize epileptogenic lesions (82% [74-88]), and predict post-surgical seizure-freedom (83% [78-87]). Overall, a high risk of bias remains in the literature; participant bias remained high across all outcomes (64-87%), as well as predictor (88-100%) and analysis (83-100%) bias. Outcome bias was low only for AI/ML studies predicting post-surgical outcomes. Conclusion. Our results support promising accuracy of AI/ML models in epilepsy diagnostics and prognostics but remain highly susceptible to bias in participants, predictors, outcome, and analysis domains which limits current translation to routine clinical practice. We encourage closer interdisciplinary collaboration between clinical and scientific groups to improve validation studies based on thorough study design, analysis, and reporting.

Elevated Thalamic Blood Flow in Self-Limited Epilepsy with Centrotemporal Spikes
Children with self-limited epilepsy syndrome with centrotemporal spikes (SeLECTS) exhibit altered thalamocortical connectivity, but whether thalamic function itself is abnormal remains unclear. We investigated whether thalamic blood flow, a marker of metabolism, differs between children with SeLECTS and controls, and examined the effects of spike distribution and antiseizure medications (ASMs) on thalamic perfusion. In this retrospective cohort study, we identified consecutive children with SeLECTS who underwent magnetic resonance imaging (MRI) for epilepsy evaluation (n = 44) and age- and sex-matched children who underwent MRI for non-epilepsy indications (n = 35). We quantified thalamic blood flow via manual segmentation of cerebral blood flow (CBF) sequences obtained from arterial spin labeling MRI. Clinical variables including sedation use during MRI, daily ASM use, and spike distribution (unilateral or bilateral) were extracted from medical records. Children with SeLECTS demonstrated elevated thalamic blood flow compared to controls, with the most pronounced differences in specific subgroups. Children with unilateral spikes showed the highest CBF, particularly in the thalamus contralateral to spike activity. ASM use significantly modulated thalamic blood flow: children taking oxcarbazepine showed the highest CBF, while those on levetiracetam had CBF similar to controls. Unmedicated children showed intermediate elevations. These findings demonstrate that elevated thalamic blood flow may be intrinsic to SeLECTS pathophysiology, with different ASMs producing distinct neurobiological effects. The differential medication effects may relate to their clinical efficacy and provide neurobiological rationale for treatment selection in this common childhood epilepsy syndrome.

Indistinguishability domains of neural microcircuit motifs mapped through classification scores of postsynaptic spike counts
In the kinetic modeling of a chemical synapse, the reversal potential (Esyn) and conductance amplitude (gsyn) of the synapse jointly regulate the spike-induced activity of the postsynaptic neuron. We simulate two-neuron microcircuit motifs of feedforward and feedback types using random external current pulses provided only at the presynaptic neuron. For combinations of Esyn and gsyn, corresponding to an excitatory synapse, a supervised machine learning method for classification distinguishes the motifs perfectly with a score of 1.0 when using binned counts of postsynaptic spikes as the input. Strongly inhibitory combinations of these parameters result in no postsynaptic response in both types of microcircuit motifs; hence, the classifier does not improve upon a random assignment with a score of 0.5. In other domains of the parameter space spanned by Esyn and gsyn, corresponding to diminished excitation/inhibition, the classifier fails (0.5 < score < 1), indicating the challenge in identifying the nature of the synapse inferred exclusively from postsynaptic spikes. For this task, gradient boosting gives higher classification accuracies that improve with further training, compared to other classifiers explored in this study.

DUSP5 Downregulation in Nucleus Accumbens Core Correlates with Synaptic Plasticity and Cue-Induced Cocaine Reinstatement
The United States is currently facing a drug overdose epidemic, with substance use disorder (SUD) characterized by cyclical phases of drug use, withdrawal, and relapse. The nucleus accumbens core (NAcore), a brain region critical for reward and aversion behaviors, undergoes structural and functional synaptic adaptations in response to chronic drug exposure. These changes, particularly in dopamine D1 receptor-expressing medium spiny neurons (D1-MSNs), are implicated in drug-seeking behaviors and synaptic plasticity. However, the molecular mechanisms underlying these adaptations remain poorly understood. In this study, we investigate the role of dual-specificity phosphatase 5 (DUSP5), an phosphatase known to deactivate extracellular signal-regulated kinase (ERK), in cocaine-induced neuroplasticity. While prior research has linked other DUSP family members to various drugs of abuse, the specific role of DUSP5 in cocaine addiction remains unexplored. We hypothesized that lack of DUSP5 contributes to NAcore synaptic plasticity during cue-induced cocaine reinstatement. To test this, we employed a rat cocaine self-administration model, integrated molecular analyses, and mined publicly available single-cell RNA sequencing data from cocaine-treated NAcore. Our findings aim to elucidate the potential involvement of DUSP5 in cocaine-related synaptic adaptations and behavior, addressing a significant gap in the mechanistic understanding of SUD.

Prenatal Polysubstance Exposure Alters Behaviour in Zebrafish Larvae
Substance use during pregnancy has been linked to various adverse outcomes in infants, including congenital disabilities, neurodevelopmental delays, and long-term effects such as learning difficulties. An additional concern is that newborns are often exposed to multiple substances in utero. The biological consequences of such exposure remain largely unknown. Zebrafish offer an exciting alternative to fill this gap and deepen our understanding of the biological impact of prenatal multidrug exposure. We utilized zebrafish's scalability to expose embryos to some of the most commonly used substances: nicotine, alcohol, opioids, and all their possible combinations. After embryonic drug exposure, we conducted a detailed behavioural analysis across three developmental stages. Our results revealed drug-specific outcomes, including both synergistic and antagonistic effects. Furthermore, we identified distinctive effects across development, highlighting potential developmental shifts and individual differences in resilience. Overall, these findings demonstrate that prenatal polydrug exposure results in complex, stage-dependent effects, sometimes antagonistic, which cannot be predicted from single-drug outcomes. Our study emphasizes the value of zebrafish as a model for investigating polydrug interactions and provides a framework for exploring biomarkers of vulnerability and resilience in offspring.

Not a global map, but a local hash: grid cells decorrelate the representation of position and scramble long-range distance information
Grid cells in the medial entorhinal cortex construct an intriguing multiperiodic representation of space whose properties have been the subject of much theoretical speculation. Here we combine modeling with analyses of neural population data from mice and rats to show that the grid cell representation is ideally set up to decorrelate and assign easily distinguishable labels to inputs, thus acting as a pattern separation device, much like a hash function in computing rather than a global map or metric. The multiple modules of the grid cell system allow the threshold for pattern separation to be controlled. We also extend these arguments to show how grid cells can perform pattern separation in abstract and higher-dimensional spaces. This pattern separation ability may serve to enhance episodic memory in the hippocampal formation by reducing interference between similar patterns.

System-Level Computations Underlie Visual Field Heterogeneity
Visual perception varies with eccentricity and polar angle. We investigated whether and how system-level computations, which transform visual input into perception, underlie these heterogeneities. Using the equivalent noise method and perceptual template model, we estimated gain, internal noise, and nonlinearity for orientation discrimination across eccentricity (fovea, parafovea and perifovea) and around polar angle. Performance declined with eccentricity due to decreased gain and nonlinearity and increased internal noise. Observers with stronger eccentricity effects showed greater gain decrease. Only gain varied with polar angle, higher along the horizontal than vertical meridian, and lower than upper vertical meridian, paralleling performance asymmetries. This dissociation aligns with known variations in neuronal count and tuning, suggesting that neural correlations and neural noise contribute to these system-level computations. By revealing distinct system-level computations underlying the eccentricity effect and polar angle asymmetries, our findings provide a link between perceptual heterogeneity across the visual field and neural architecture.

Neuronal Activity in Orbitofrontal Cortex during Trinary Choices under Risk
Economic choice entails computing and comparing the subjective values of different goods. Orbitofrontal cortex (OFC) is thought to contribute to both operations. However, previous work focused almost exclusively on binary choices, raising the question of whether current notions hold for multinary choices. Here we recorded from rhesus monkeys making trinary choices. Offers varied on three dimensions -- juice flavor, quantity, and probability. In these experiments, quantity and probability varied continuously within a preset range. Animal choices were generally risk seeking and satisfied independence of irrelevant alternatives (IIA) -- a fundamental assumption in standard economic theory. Different neurons encoded the values of individual offers, the choice outcome, and the chosen value -- i.e., the same variables previously identified under binary choices. In addition, other cell groups encoded the chosen probability and the chosen hemifield. Notably, the activity of offer value cells reflected the risk attitude and fluctuated from session to session in ways that matched fluctuations observed behaviorally. In other words, the activity of these neurons reflected the subjective nature of value. Importantly, the representation of decision variables in OFC was invariant to changes in menu size -- a property that effectively implies IIA.