bioRxiv Subject Collection: Neuroscience's Journal

Unilateral damage to the entopeduncular nucleus causes forelimb motor dysfunction in rats
Background: The entopeduncular nucleus (EP), corresponding to the human globus pallidus internal segment, is a basal ganglia output nucleus, and plays a critical role in motor control. However, the impact of EP damage on skilled motor function and the relationship between its damage in stroke, such as internal capsule hemorrhage (ICH), and motor dysfunction remains unclear. This study aimed to clarify whether EP damage causes motor dysfunction in two disease models. Methods: EP-related motor dysfunction was investigated by inducing localized unilateral EP damage in Long-Evans rats using a stereotactic kainic acid (KA) injection. Motor function was assessed using a single-pellet reaching task pre-injection and on postoperative days 2, 7, 14, 21, and 28. Immunohistochemical staining for NeuN, somatostatin (SST), and parvalbumin was conducted to quantify damage and its correlation with motor outcomes. In addition, unilateral ICH was induced via stereotactic injection of collagenase type IV, which dissolves the vascular basement membrane, into the internal capsule (IC) of Long-Evans rats. Injury sites were classified into the IC, dorsomedial region from the IC, ventral lateral region from the IC, and EP, and their volumes were measured. Measured volumes were analyzed for correlations with motor function assessments. Results: KA-induced EP damage significantly reduced reaching success rates on postoperative day 2 compared to those in the control group (p<0.05). Immunohistochemical analysis showed that reaching success rates on day 28 positively correlated with the numbers of remaining NeuN-positive and SST-positive neurons (p<0.05). In the ICH experiment, all rats significantly reduced the success rate of the reaching task to 0% on day 2, and the success rate on day 28 correlated positively with the remaining EP volume, but not with total lesion volume. Conclusions: EP damage was strongly associated with motor impairments, highlighting its critical role in motor control and recovery.

Radionuclide selection influences imaging outcomes in immunoPET with a brain-penetrant anti-Aβ antibody
Background: Bispecific antibodies exploiting receptor-mediated transcytosis offer a promising strategy to overcome limited blood-brain barrier permeability in Alzheimer's disease (AD) therapy and imaging. Lecanemab-Fab8D3 (Lec-Fab8D3), a bispecific anti-amyloid beta (A{beta}) antibody engineered for enhanced brain delivery, holds potential as a companion immunoPET imaging diagnostic with the novel lecanemab immunotherapy. This study aimed to compare three radionuclides--zirconium-89 (89Zr), copper-64 (64Cu), and iodine-124 (124I)--for PET imaging with Lec-Fab8D3 to study its in vivo brain distribution and evaluate its potential as an AD companion diagnostic. Methods: Lec-Fab8D3 was conjugated to DFO* or NODAGA for 89Zr and 64Cu radiolabeling, respectively, or directly radioiodinated with 124I. PET imaging was performed in the Tg-ArcSwe mouse model of A{beta} pathology and wild-type (WT) littermates at multiple time points post administration of the radiolabeled antibody, followed by ex vivo biodistribution, autoradiography, and A{beta} quantification to assess brain uptake, specificity, and distribution of the radiolabeled Lec-Fab8D3. Results: Radiolabeled Lec-Fab8D3 variants showed retained binding properties with high radiochemical purity and yields. PET imaging demonstrated cortical brain uptake of all three tradiotracers in Tg-ArcSwe mice, with [89Zr]Zr-DFO*-Lec-Fab8D3 and [124I]I-Lec-Fab8D3 showing the best discrimination between Tg-ArcSwe and WT mice at 48-72 h post-injection. The highest absolute brain retention, combined with a lower brain-to-cerebellum ratio, was observed in both Tg-ArcSwe and WT mice that received the radiometal-labeled (89Zr and 64Cu) antibody, likely due to the residualizing nature of radiometals. Ex vivo analyses confirmed PET findings, and immunostaining demonstrated co-localization of Lec-Fab8D3 with A{beta} deposits. Conclusions: ImmunoPET imaging with bispecific Lec-Fab8D3 enables specific detection of brain A{beta} pathology in an AD mouse model. 89Zr was superior to 64Cu due to a more compatible half-life, while 124I displayed higher regional contrast than both radiometals, despite lower overall brain signal. The combined findings from radiometal- and iodine-based immunoPET will enhance our understanding of intra-brain distribution of bispecific antibodies. Furthermore, this highlights the importance of the choice of radiolabeling strategy and how it will impact the outcome of immunoPET with bispecific A{beta} antibodies.

Lineage-Traced PTBP1 Depletion in Mature Astrocytes Reveals Distinct Splicing Alterations Without Neuronal Features
Astrocyte-to-neuron reprogramming via depletion of PTBP1, a potent repressor of neuronal splicing, has been proposed as a therapeutic strategy, but its efficacy remains debated. While some reported successful conversion, others disputed this, citing a lack of neuronal gene expression as evidence of failed reprogramming. This interpretation was further challenged, attributed to incomplete PTBP1 inactivation, fueling ongoing controversy. Mechanistic understanding of the conversion, or the lack thereof, requires investigating, in conjunction with lineage tracing, the effect of Ptbp1 loss of function in mature astrocytes on RNA splicing, which has not yet been examined. Here, we genetically ablated PTBP1 in adult Aldh1l1-Cre/ERT2 Ai14 mice to determine whether lineage traced Ptbp1 knockout astrocytes exhibited RNA splicing alterations congruent with neuronal differentiation. We found no widespread induction of neurons, despite a minuscule fraction of knockout cells showing neuron-like transcriptomic signatures. Importantly, PTBP1 loss in mature astrocytes induced splicing alterations unlike neuronal splicing patterns. These findings suggest that targeting PTBP1 alone is ineffective to drive neuronal reprogramming and highlight the need for combining splicing and lineage analyses. Loss of astrocytic PTBP1 is insufficient to induce neuronal splicing, contrasting with its well-known role in other non-neuronal cells, and instead affects a distinct astrocytic splicing program.

Selective vulnerability of cerebral vasculature to NOTCH3 variants in small vessel disease and rescue by phosphodiesterase-5 inhibitor
NOTCH3 variants cause CADASIL the most common genetic form of small vessel disease (SVD) and vascular dementia (VaD) and increase the stroke and SVD/VaD risk. CADASIL is a systemic vasculopathy but predominantly manifests in the brain. The molecular mechanisms of CADASIL remain largely unclear with no specific available treatments. NOTCH3 is primarily expressed in vascular smooth muscle cells (VSMCs). Using human induced pluripotent stem cell (iPSC) models and developmental lineage-specific VSMC differentiation, we revealed a selective vulnerability of cerebral but not peripheral VSMC mimics to NOTCH3 variants. Transcriptomic, protein and functional analyses demonstrated a switching of CADASIL iPSC-VSMCs from a contractile to a synthetic phenotype, accompanied with extensive extracellular matrix accumulation and impairment of cell adhesion leading to anoikis. Importantly, we describe an endothelial independent nitric oxide signalling in VSMCs which was dysregulated in the CADASIL iPSC-VSMCs, and posphodiesterase-5 (PDE5) inhibition successfully rescued the abnormal VSMC function, suggesting a novel therapeutic strategy. Our findings offer new mechanistic insights into brain specific phenotype in NOTCH3-associated SVD/VaD and support patient-specific iPSCs to be a valuable model for identifying targeted treatment for NOTCH3-associated SVD/VaD.

Deciphering the combinatorial expression pattern and genetic regulatory mechanisms of Beats and Sides in the olfactory circuits of Drosophila
Over the past decades, many critical molecular players have been uncovered to control distinct steps in olfactory circuit assembly in Drosophila. Among these, multi-member gene families of cell surface proteins are of interest because they can act as neuron-specific identification/recognition tags in combinations and contribute to circuit assembly in complex brains through their heterophilic or homophilic interactions. Recently, a multi-protein interactome has been described between the Beat and Side families of IgSF proteins. Here, we use the publicly available single-cell RNA-seq datasets and newly generated gene trap transgenic driver lines to probe the in vivo spatial expression pattern of the beat/side gene families in odorant receptor neurons (ORNs) and their synaptic target projection neurons (PNs). Our results revealed that each ORN and its synaptic target PN class expresses a class-specific combination of beat/side genes, hierarchically regulated by lineage-specific genetic programs. Though ORNs or PNs from closer lineages tend to possess more similar beat/side profiles, we also found many examples of divergence from this pattern among closely related ORNs and closely related PNs. To explore whether the class-specific combination of beats/sides defines ORN-PN matching specificity, we perturbed presynaptic beat-IIa and postsynaptic side-IV in two ORN-PN partners. However, disruption of Beat-IIa-Side-IV interaction did not produce any significant mistargeting in these two examined glomeruli. Though without affecting general glomerular targeting, knockdown of side in ORNs leads to the reduction of synaptic development. Interestingly, we found conserved expression patterns of beat/side orthologs across ORNs in ants and mosquitoes, indicating the shared regulatory strategies specifying the expression of these duplicated paralogs in insect evolution. Overall, this comprehensive analysis of expression patterns lays a foundation for in-depth functional investigations into how Beat/Side combinatorial expression contributes to olfactory circuit assembly.

γ-glutamyl-cysteine is a critical intermediate in glutathione-led amelioration of PFOS-neurotoxicity
Per/polyfluoroalkyl substances (PFAS) are used in a variety of industrial and consumer applications due to their distinctive properties. Nonetheless, these substances are ubiquitous and pose significant risks to the environment, wildlife, and human health. Perfluorooctane sulfonate (PFOS), which was previously one of the most commonly utilized PFAS, has prolonged half-life in both humans and the environment. Despite the cessation of its production, PFOS remains one of the prevalent contaminants amongst PFAS. Recent research demonstrates that PFOS is relatively resistant to elimination from the human body and accumulates to a greater extent in older adults. PFOS has been shown to affect the nervous system and its functions, although the intricate mechanisms underlying its toxicity remain largely obscure. Earlier studies utilizing Caenorhabditis elegans indicated that dopaminergic neurons are particularly vulnerable to PFOS neurotoxicity, with glutathione (GSH) playing a role in mitigating neurodegeneration. Curiously, none of the antioxidant treatments evaluated, including N-acetyl-cysteine, produced favorable outcomes, despite N-acetyl-cysteine being a precursor to GSH. This study investigates the GSH synthesis pathway to elucidate critical mechanisms. We assessed the effects of GSH precursors and intermediates on PFOS neurotoxicity. The GSH precursors, cysteine, glutamate, and the combination of cysteine and glutamate did not demonstrate beneficial effects. However, the crucial GSH synthesis intermediate, {gamma}-glutamyl-cysteine, provided neuroprotection comparable to that of GSH. Notably, no changes were observed at the transcriptomic or proteomic levels of GSH synthesis enzymes in C. elegans and SH-SY5Y cells, respectively. This study effectively uncovers a novel mechanism that addresses existing knowledge gaps pertaining to PFOS neurotoxicity.

Molecular Subtyping Based on Hippocampal Cryptic Exon Burden Reveals Proteome-wide Changes Associated with TDP-43 Pathology across the Spectrum of LATE and Alzheimer's Disease
TDP-43 pathology is a defining feature of Limbic-Predominant Age-Related TDP-43 Encephalopathy neuropathologic change (LATE-NC) and is frequently comorbid with Alzheimer's disease neuropathologic change (ADNC). However, the molecular consequences of co-occurring LATE-NC and ADNC pathology (TDP-43, {beta}-amyloid, and tau protein pathologies) remain unclear. Here, we conducted a comparative biochemical, molecular, and proteomic analysis of hippocampal tissue from 90 individuals spanning control, LATE-NC, ADNC, and ADNC+LATE-NC groups to assess the impact of cryptic exon (CE) inclusion, phosphorylated TDP-43 pathology (pTDP-43), and AD-related pathologies ({beta}-amyloid, and tau) on the proteome. ADNC+LATE-NC cases exhibited the highest burden of CE inclusion as quantified by measuring the levels of known TDP-43 regulated CEs within eight transcripts: STMN2, UNC13A, ELAVL3, KALRN, ARHGAP32, CAMK2B, PFKP, and SYT7. While CE levels correlated with pTDP-43 pathology, they were more strongly correlated with each other, suggesting that the molecular signature of CE inclusion may serve as a more sensitive measure of TDP-43 dysfunction than pTDP-43 pathology alone. Unbiased classification based on the relative abundance of these eight CEs stratified individual cases into low, intermediate, and high CE burden subtypes, largely independent of {beta}-amyloid and tau pathology. Proteome-wide correlation analysis revealed a bias toward reduced protein levels from genes harboring TDP-43-regulated CEs in cases with high cumulative CE burden. Notably, proteins significantly decreased under high CE burden included canonical STMN2, ELAVL3, and KALRN, as well as kinesin proteins that are genetically associated with amyotrophic lateral sclerosis. Co-expression network analysis identified both shared and distinct biological processes across CE subtypes and pathways associated with pTDP-43, tau, {beta}-amyloid pathologies, and CE accumulation in the hippocampus. Protein modules associated with TDP-43 loss of function were prioritized by integrating proteomic data from TDP-43 depleted human neurons with the hippocampal co-expression network. Specifically, we observed decreased endosomal vesicle, microtubule-binding, and synaptic modules, alongside an increase in RNA-binding modules. These results provide new insights into the proteomic impact of CE burden across the spectrum of LATE and AD pathological severity, highlighting the molecular consequences of TDP-43 dysfunction in neurodegenerative disease.

FastACI: A Toolbox for Investigating Auditory Perception using Reverse Correlation
The fastACI toolbox provides a compilation of tools for collecting and analyzing data from auditory reverse-correlation experiments. These experiments involve behavioral listening tasks including one or more target sounds presented with some random fluctuation, typically in the form of additive background noise. In turn, the paired stimulus-response data from each trial can be used to assess the relevant acoustic features that were effectively used by the listener while performing the task. The results are summarized as a matrix of perceptual weights termed auditory classification image (ACI). The framework provided by the toolbox is flexible and it has been so far used to probe different auditory mechanisms such as tone-in-noise detection, amplitude modulation detection, phoneme-in-noise categorization, and word segmentation. In this article, we present the structure of the toolbox, how it can be used to run existing experiments or design new ones, as well as the main options for analyzing the collected data. We then illustrate the capabilities of the toolbox through five case studies: a replication of a pioneering reverse correlation study from 1975, an example of reproduction of the analyses of one of our previous studies, a comparison of the results of three phoneme-categorization experiments, and a quantification of how noise type and estimation method affect the quality of the resulting auditory classification image.

A theoretical mechanism for ocular emmetropization.
Currently, the human myopia epidemic is estimated to affect nearly 3 billion persons, yet experimental refractive error research is still hampered by the lack of a plausible theoretical mechanism explaining how the eye detects defocus and grows to minimise error (emmetropization). Applied lens defocus, either positive or negative, to animals including human, induces rapid changes in both refraction and axial length. Such changes have been linked to the rate of transfer of fluid from the vitreous chamber across the Retinal Pigment Epithelium (RPE). However, the theoretical basis of sensing the sign of defocus still eludes, as does the physiological operationalisation of this defocus signal. We propose that the signal for the sign of defocus is contained within the pattern of temporal modulation of light on the retina during the myriad saccadic eye movements performed every minute. This pattern is generated by a combination of the ocular point spread function for the defocused eye, narrow cone photoreceptor acceptance of light and the modification of photoreceptor directionality by the retinal shear that accompanies saccadic eye movements. Thus, under conditions of defocus, performing a saccade across a simple visual grating produces an out of focus dynamic stimulus pattern on the retina which is sawtooth-like, the profile being dependent on the sign of defocus. Positive lens defocus induces a fast-OFF/slow-ON sawtooth-like luminance modulation transfer function, while negative lens defocus results in a fast-ON/slow-OFF temporal pattern. Such patterns produce relative incremental or decremental changes in the trans-epithelial electrical potential and consequent changes in fluid absorption across the RPE. Thus, rearing with positive lens defocus is associated with an increase in trans-epithelial potential, increasing trans-epithelial absorption and decreasing vitreous chamber size, while rearing with a negative lens decreases the trans-epithelial electrical potential and decreases RPE fluid absorption, increasing eyeball size. In both cases, retinal images become more in-focus, and the refractive error tends to zero. The resultant is emmetropization, modifying eyeball size in response to applied defocus in a stable negative feedback fashion.

Sleep loss suppresses the motivation to prepare for sleep
In the period preceding sleep, humans and other animals display a stereotyped repertoire of behaviors-including hygiene-related activities and preparing a place to sleep. Evidence suggests that this pre-sleep phase actively contributes to sleep initiation and quality. Nonetheless, individuals can sometimes fall asleep without preparation, even under undesirable circumstances. These abrupt transitions into sleep can have severe consequences, particularly in high-risk environments. Although progress has been made in identifying neuronal populations controlling sleep-wake states and mechanisms regulating cortical oscillations during sleep, little is known about the natural processes that govern the pre-sleep phase, under baseline conditions and following sleep loss. Here, we examine factors regulating pre-sleep behaviors using environmental and behavioral manipulations, video recordings, machine-learning-based tracking, and EEG-EMG analysis in freely-behaving mice. We focus on nest-building, a key pre-sleep behavior, and assess its modulation by time of day and sleep deprivation. We find that mice are highly motivated to build nests during the light phase but show reduced motivation during most of the dark phase. Sleep deprivation significantly suppresses pre-sleep nest-building and promotes the direct initiation of sleep. Varying amounts of sleep deprivation, from 2-6 hours, uniformly suppress nest-building. This suppression is not due to stress, as mice exposed to acute restraint stress engage robustly in nest-building. Our findings provide insight into processes regulating the transition from wakefulness to sleep. Understanding pre-sleep regulation has important implications for treating sleep-onset difficulties-prevalent in insomnia and predictive of cognitive decline, and for mitigating risks associated with uncontrolled sleep onset in high-stakes situations.

Power-to-power cross-frequency coupling as a novel approach for temporal lobe seizure detection and analysis
Objective: Power-to-power cross-frequency coupling (CFC) is a novel method to index the dynamic spatio-temporal interactions between brain rhythms, including high frequency oscillations (HFOs). This research evaluates this promising method's capacity for seizure detection with intracranial EEG. Seizures can be conceptualized as composites of different electrographic patterns including (1) spike, (2) ripple-on-spike, and (3) ripple-on-oscillation. This study also performs a basic CFC analysis of each of these components which has potential to further the understanding of epileptogenic processes. Methods: In this study, deep learning networks including Stacked Sparse Autoencoder (SSAE) and Long Short Term Memory (LSTM) are trained to detect seizures and help characterize CFC patterns for these three common seizure components. The analysis uses intracranial EEG (iEEG) records from the ieeg.org (Mayo Clinic files) database. Temporal Lobe Epilepsy (TLE) seizures (n = 120) from 26 patients were analyzed along with segments of background activity. Power-to-power coupling was calculated between all frequencies 1-250 Hz pairwise using the EEGLAB toolbox. CFC matrices of seizure and background activity were used as training or testing inputs to the autoencoder. Results: The trained network was able to recognize background and seizure segments (not used in training) with a sensitivity of 90.2%, specificity of 96.8% and overall accuracy of 93.4%. The three seizure components (spike, ripple-on-spike, ripple-on-oscillation) were also observed to have unique CFC signatures. Conclusions: The results provide evidence both for (1) the relevance of power-to-power coupling (PPC) for TLE seizure detection in iEEG, as well as (2) there existing unique PPC signatures of three common seizure components.

Social experience alters behaviors by reprogramming the Fruitless pathway and circadian state in Drosophila
Animals thrive with social interactions and suffer significant adverse mental and physical health consequences when isolated. From flies to humans, social experience affects various cognitive and behavioral processes. Here, we use the fruit flies to show that group housing suppresses male courtship vigor as a result of a decrease in the evoked neural responses in courtship circuits. Bulk tissue RNAseq and single-cell RNAseq from fru- and dsx-positive cells from grouped or isolated male brains revealed that social isolation increases the number of fru- and dsx-positive neurons, elevates fru transcript levels in specific clock neurons, and dsx transcript levels throughout the brain. Knocking down fruM in the fru-positive neurons in brains decreases courtship in isolated males to levels comparable to those of group-housed males. Furthermore, group housing increases the expression of sr and Hr38 genes encoding neural activity-induced transcription factors in most neurons within social circuits. We found that knocking down sr in fru positive neurons effectively eliminates the impact of social experience by increasing courtship in group-housed males. Importantly, social experience also alters the expression of FruM/DsxM target genes regulating circadian states throughout the brain. Disrupting circadian gene function also diminishes the effect of group housing on courtship. Our findings suggest that group housing/social enrichment suppress courtship by reprogramming the circadian arousal state, whereas courtship-elevating effects of social isolation rely on changes in FruM expression and function. These results are significant as they point to modulation of circadian arousal state as a possible central strategy for mediating the pleiotropic effects of social experience on organismal responses.

Non-Physiological Potassium Concentrations in Commercial Culture Media Trigger Acute Epileptiform Activity in Human iPSC-Derived Neurons
Neuronal in vitro cultures are pivotal for studying brain electrophysiological function and dysfunction. Neuronal activity and communication are regulated by extracellular ion concentrations. Therefore, cell culture medium ion concentrations should ideally mimic those of cerebrospinal fluid (CSF) - considered as the milieu for brain cells in vivo. In this study, we demonstrate that commonly used cell culture media, including Neurobasal (+/- A), Neurobasal Plus, and BrainPhys media, do not accurately replicate human CSF ion concentrations. Using human iPSC-derived neuronal networks on microelectrode arrays, we show that the abnormally high potassium concentrations present in all tested cell culture media induce acute epileptiform activity, similar to that elicited by the convulsive drug 4-aminopyridine. These findings raise a critical question: How can human in vitro neuronal activity be defined as physiological and reliably distinguished from pathophysiological activity, if the routinely used ion concentrations in in vitro experiments are causing aberrant neuronal activity?

The Neuromuscular system of Chironomus vitellinus (Diptera: Chironomidae)
Chironomids are important laboratory model organisms used to assess toxicity in freshwater environments. Cell and tissue features are not commonly used as chironomid markers to detect toxicity, but they could be extremely helpful in identifying acute and chronic effects of pollutants. The nervous system is an excellent cellular candidate since it is reactive to toxic substances. However, a detailed description of the chironomid nervous system is required prior to considering it as a candidate for a cellular toxicity marker. The present study describes the central ganglia, nerves, axons, and the neuromuscular system of Chironomus vitellinus (Freeman, 1961) to facilitate its use as a model organism in environmental studies. We find that the structure of the C. vitellinus central nervous system is identical to that observed in other Chironomus larvae. We then focused our study on the first abdominal segment and labeled the 31 hemi-segmental muscles according to a nomenclature based on their position and orientation. We also characterized their innervation and assigned the nerves a nomenclature based on their terminals location in the muscle tissue. Finally, we investigated the neuromuscular junctions (NMJs) throughout this segment and defined four types of NMJs illustrating their great variability in size and shape. We selected a model NMJ, VEL 2, and quantified its mean bouton number and muscle size. Together with documenting a neurobiological system that could be informative to insects comparative biology, these results could help establish the Chironomus NMJ as an aquatic toxicity marker.

How individual vigor shapes human-human physical interaction
The speed of voluntary movements varies systematically, with some individuals moving consistently faster than others across different actions. These variations, conceptualized as vigor, reflect a time-effort-accuracy tradeoff in motor planning. How do two mechanically coupled partners with different individual vigors collaborate, e.g. to move a table together? Here, we show that such dyads coordinate goal-directed movements with minimal interaction force, exhibiting a dyadic vigor with similar characteristics as individual vigor. The emerging dyadic motor plan is strongly influenced by the slower partner, whose vigor predicts dyadic vigor, with effects lasting beyond connected practice. Computational modeling reveals the critical role of the partners' movement timing's uncertainty and vigor in shaping coordination, allowing to predict dyadic movements from individual behavior across diverse conditions. These findings shed light on the mechanisms underlying successful human collaboration, and may be used in applications ranging from physical training and rehabilitation to collaborative robotics for manufacturing.

Non-conscious Multisensory Integration in the Ventriloquist Effect
The degree to which information from distinct sensory modalities can interact in the absence of conscious awareness remains controversial. According to the Global Neuronal Workspace Theory (GNWT), unconscious sensory information remains relatively confined to sensory cortex and should not be capable of interacting with other modalities until it is broadcast into the (conscious) global workspace comprising late (>300ms) frontal-parietal activation. The ventriloquist effect is a classical multisensory integration phenomenon that refers to the misperception of sound location towards concurrent visual stimulation, such as perceiving the voice of a ventriloquist actor as coming from the moving dummy. Here we used meta-contrast masking to render a brief flash stimulus non-conscious while participants performed a sound localization task. We found that, despite being at chance performance in discriminating the flash location, participants were nevertheless biased to localize sounds towards the unconscious flash locations. The effect was present in virtually all participants, was nearly as large as the effect on conscious trials, and was robust to controls for individual differences in task performance. Decoding analyses of concurrently recorded EEG signals showed that the non-conscious flash location information was present up until around 230ms but not after; confirming that the visual influence on sound perception likely occurred before conscious broadcast. Our findings suggest that subjective perception is not required for the integration of signals originating in distinct sensory modalities prompting new questions about the role of subjective perception in multisensory integration.

Bazedoxifene rescues sexually dimorphic autistic-like abnormalities in mice carrying a biallelic MDGA1 mutation
In the accompanying study, we describe two pairs of MDGA1 missense mutations (Val116Met/Ala688Val and Tyr635Cys/Glu756Gln) from two patients with autism spectrum disorders (ASDs) and how these mutations exert distinct abnormalities in biological processes during central nervous system development. Here, we generated knockin (KI) mice harboring the murine version of Tyr635Cys/Glu756Gln MDGA1 (Mdga1Y636C/E751Q) and performed extensive behavioral analyses. Male KI mouse pups and adults exhibited impaired ultrasonic vocalizations and sensorimotor gating (ASD-relevant behavioral deficits), and thus exhibited phenotypes different from those of male Mdga1 conditional knockout (cKO) mice. In contrast, adult female KI mice did not exhibit a range of ASD-like behavioral abnormalities. Electrophysiological analyses performed using both juvenile and adult Mdga1Y636C/E751Q mice revealed sexually dimorphic and developmental stage-dependent compromises of GABAergic synaptic inhibition. In addition, proteomics analyses showed that the phospho-proteomic phenotypes differed between Mdga1Y636C/E751Q and Mdga1-cKO mice. Treatment of male Mdga1Y636C/E751Q KI mice with the FDA-approved estrogen receptor modulator, bazedoxifene, restores the KI-related molecular, GABAergic synapse, and behavioral changes. Collectively, our results suggest a novel treatment strategy for ASDs that present developmentally regulated and/or sexually dimorphic features.

A characterization of mouse retinal ganglion cell types labeled with AAV tools
The mouse retina is made-up of approximately 150 types of neurons each with unique characteristics and functions in interpreting visual information. Recent efforts to categorize cell types using molecular markers, morphology, and electrophysiological response properties have provided a wealth of information and a host of tools for studying specific cell types. AAV-based approaches have several advantages over transgenic mouse lines, including ease of application to many different animal models without extensive crossing and their amenability to intersectional approaches. Here, we provide an in-depth characterization of retinal ganglion cell types labeled by two AAV vectors drawn from a recent panel of constructs with synthetic promoters. Each promoter analyzed here was derived from a gene expressed in a cell type specific manner. Using a combination of morphology, molecular markers, and electrophysiological measurement of light responses, we found that each vector labeled distinct subsets of RGCs. However, both labeled more cell types than expected from the expression pattern of the promoters endogenous gene. We then characterized the projection patterns of these RGC types to the brain, finding that each AAV type labeled distinct axonal populations. These tools provide new access to a unique subset of cells and will be instrumental to future studies analyzing their functions and connectivity.

Direct Binding of FGFR3 Autoantibodies to Sensory Neurons Drives Hyperexcitability and Mechanical Hypersensitivity
Sensory neuronopathies (SNN) and small fiber neuropathies (SFN) are debilitating disorders often associated with neuropathic pain, yet their underlying mechanisms remain poorly understood. Autoantibodies against fibroblast growth factor receptor 3 (-FGFR3) have been identified in a subset of patients, but their pathological significance has not been established. Here, we describe that -FGFR3-positive patients consistently report neuropathic pain and display a distinct clinical phenotype characterized by large-fiber involvement and non-length-dependent fiber loss, suggesting dorsal root ganglia (DRG) dysfunction. We demonstrate that -FGFR3 bind to sensory neurons within dorsal root ganglia (DRG). We validated both at the transcript and protein level that the target of autoantibodies, FGFR3, is expressed in human sensory neurons and that therefore -FGFR3 could find their target in primary afferents. DRG neurons exposed to -FGFR3 rapidly acquired a hyperexcitability phenotype. Injection of -FGFR3-positive patient serum in rats caused mechanical hypersensitivity, mirroring patient-reported pain symptoms. Mechanistically, -FGFR3 activated the Mitogen-activated protein kinases (MAPK) signaling cascade, specifically extracellular signal-regulated kinase (ERK) and p38, which are known to enhance neuronal excitability. Epitope mapping revealed key extracellular epitopes on FGFR3. Blocking these epitopes prevented -FGFR3-induced sensory neuron hyperexcitability, thus showing that the autoantibody binding of the FGFR3 extracellular domain is a key factor affecting DRG neurons. Our work suggests that beyond their role as biomarkers, -FGFR3 actively contribute to pain hypersensitivity by acting on the DRG. This positions both -FGFR3 and FGFR3 signaling as a potential therapeutic targets for modulating sensory neuron excitability and treating autoimmune neuropathies

Measurements and simulations of transmembrane water exchange by diffusion NMR methods: From yeast cells to optic nerve ex vivo
Non-invasive measurements of exchange is paramount in different fields, ranging from material to biological sciences, and may even blur micro-structural or other characteristics of multi compartmental systems studied by MR methods. Despite the growing interest in diffusion-exchange studies of complex systems--where at least one exchanging component exhibits non-Gaussian diffusion--comparative studies remain scarce. Most existing investigations have applied different diffusion MR methods to different biological samples under varying experimental conditions, making direct comparisons difficult. Moreover, the lack of a gold standard for exchange rate measurements further complicates efforts to validate and interpret results. To address these challenges, we employed two diffusion NMR-based methods--the constant-gradient pulsed field gradient (CG-PFG) and the recently introduced filter-exchange NMR spectroscopy (FEXSY)--to investigate apparent water exchange in yeast cells and optic nerves, both before and after fixation. We first evaluated the effect of the q-values on the extracted indices and then evaluated the stability and reproducibility of the measurements. The CG-PFG and FEXSY experiments were collected on the same sample to allow for comparison of the results. The intracellular mean residence times (MRTs) ({tau}i) extracted from the log-linear fit of the CG-PFG NMR experiments were found to be 554{+/-}6ms and 337{+/-}10ms for yeast cells before and after fixation, respectively. The respective {tau}i values extracted from the FEXSY experiments before and after fixation were found to be 368{+/-}14ms and 146{+/-}24ms, respectively. Despite the difference in absolute values of the MRTs, the same qualitative behavior is observed in the two methodologies, and both could be analyzed using the bi-compartmental Karger model. The same methodologies were then used to study exchange in the more complex porcine optic nerves. There, the bi-compartmental Karger model analysis is shown to be inadequate. Extensive Monte Carlo simulations are used to narrow down on the most possible explanation, suggesting that optic nerves are multi-compartmental systems where not all spins are free to undergo exchange. Supporting theoretical calculation point to the existence of at least one additional non-exchanging restricted compartment. Thus, a tri-compartmental model is derived and used to analyze the data. The new model fits the data significantly better and results in dramatically different exchange rates when used on white matter (WM) data: CG-PFG experiments were found to be 730{+/-}40ms and 803{+/-}16ms for optic nerve before and after fixation, respectively. The respective {tau}i values extracted from the FEXSY experiments before and after fixation were found to be 530{+/-}125ms and 387{+/-}104ms, respectively. These values are considerably smaller than values that were previously reported. Finally, we use simulations to show that the quantitative discrepancy between the CG-PFG and FEXSY can be attributed, at least partially, to the difference in T2 values between the intra- and extracellular compartments. We thus encourage the pairing of exchange and spin-spin relaxation measurements in future works. We end with a discussion on the current state of the diffusion-exchange, and in an attempt to put a spotlight on essential corner stones that are still missing despite the great advance of recent years -- experimental standardization, method comparison and adequate modeling.

Neuronal avalanches as a predictive biomarker of BCI performance- towards a tool to guide tailored training
Brain-Computer Interfaces (BCIs) based on motor imagery (MI) hold promise for restoring control in individuals with motor impairments. However, up to 30% of users remain unable to effectively use BCIs: a phenomenon termed BCI inefficiency. This study addresses a major limitation in current BCI training protocols: the use of fixed-length training paradigms that ignore individual learning variability. We propose a novel approach that leverages neuronal avalanches, spatiotemporal cascades of brain activity, as biomarkers to characterize and predict user-specific learning mechanism. Using electroencephalography (EEG) data collected across four MI-BCI training sessions in 20 healthy participants, we extracted two features: avalanche length and activations. These features revealed significant training and task-condition effects, particularly in later sessions. Crucially, changes in these features across sessions (avalanche length and activations) correlated significantly with BCI performance and enabled prediction of future BCI success via longitudinal Support Vector Regression and Classification models. Predictive accuracy reached up to 91%, with notable improvements after spatial filtering based on selected regions of interest. These findings demonstrate the utility of neuronal avalanche dynamics as robust biomarkers for BCI training, supporting the development of personalized protocols aimed at mitigating BCI illiteracy.

A genetically encoded fluorescent sensor for monitoring spatiotemporal prostaglandin E2 dynamics in vivo
Prostaglandin E2 (PGE2) is an important lipid signaling molecule that regulates a wide range of physiological and pathological processes. However, its dynamics during these processes are largely unknown due to the lack of tools to directly visualize PGE2 with high spatiotemporal resolution. Here, we developed and characterized a genetically encoded PGE2 sensor, which we call GRABPGE2-1.0 (PGE2-1.0), that has high specificity for PGE2, nanomolar affinity, rapid kinetics, and high spatial resolution when expressed both in vitro and in vivo. Using fiber-photometry recordings, we found that PGE2-1.0 can reliably monitor endogenous PGE2 dynamics in the preoptic area in the brain during acute inflammation. The wide-field in vivo imaging with PGE2-1.0 reveals spatial heterogeneity in cortex-wide PGE2 dynamics during acute inflammation and seizure. Thus, our PGE2-1.0 sensor can be used to detect endogenous PGE2 dynamics with high spatiotemporal resolution, providing a robust tool for studying PGE2 under specific physiological and pathological conditions.

Novel mouse model reveals neurodevelopmental origin of PMM2-CDG brain pathology
Congenital disorders of glycosylation (CDG) are a group of neurogenetic conditions resulting from disruptions in the cellular glycosylation machinery. The majority of CDG patients have compound heterozygous pathogenic variants in the phosphomannomutase 2 (PMM2) gene. Individuals with PMM2-CDG exhibit multi-systemic symptoms, prominently featuring neurological deficits with nearly all patients exhibiting cerebellar hypoplasia and ataxia. To overcome embryonic lethality caused by whole body knock-out of Pmm2 and mimic patient-related compound heterozygous pathogenic variants, we paired a Pmm2 flox allele (Pmm2fl) with a catalytically inactive knock-in allele (Pmm2R137H), commonly present in PMM2-CDG patients. Mice with post-mitotic loss of PMM2 from neurons or astrocytes are indistinguishable from unaffected littermates, including in a broad battery of neurological assessments. In contrast, removal of PMM2 from embryonic neural precursor cells leads to cerebellar hypoplasia, ataxia, seizures, and early lethality. Comprehensive multi-omics profiling, including metabolomics, glycomics, single-cell transcriptomics, proteomics, and glycoproteomics, reveal widespread molecular disturbances throughout the brain, with the cerebellum showing the most pronounced disruption. These findings highlight the heightened dependency of the developing cerebellum on intact N-glycosylation, aligning with clinical observations in PMM2-CDG patients. Importantly, glycoproteomic alterations identified in our mouse model are corroborated in PMM2-CDG patient post-mortem cerebellar tissue, underscoring the translational relevance of our findings and implicating impaired synaptic transmission as a key pathogenic mechanism.

Paraventricular Thalamus Hyperactivity Mediates Stress-Induced Sensitization of Unlearned Fear but Not Stress-Enhanced Fear Learning (SEFL)
Exposure to stress can cause long-lasting enhancement of fear and other defensive responses that extend beyond the cues or contexts associated with the original traumatic event. These nonassociative consequences of stress, referred to as fear sensitization, are thought to underlie some symptoms of trauma-related disorders. Fear sensitization has been predominately studied using the Stress-Enhanced Fear Learning (SEFL) paradigm, which models the stress-induced amplification of fear learning. Less is known about the mechanisms through which unlearned fear responses are sensitized by stress. Here, we investigated the neural mechanisms for sensitization of unlearned fear responses using a paradigm we termed Stress-Enhanced Fear Responding (SEFR). In this model, mice exposed to a single session of footshock stress exhibit enhanced freezing to a novel tone stimulus. To investigate brain regions that might mediate SEFR, we first used c-Fos mapping to identify neural activity changes associated with stress-induced enhancement of unlearned fear. Our c-Fos screen identified the posterior paraventricular thalamus (pPVT) as a region that was persistently hyperactive after footshock stress and whose activity correlated with behavioral expression of SEFR. Using fiber photometry, we observed that SEFR, but not SEFL, was associated with increased activity in the pPVT. Next, we found that chemogenetic inhibition of the pPVT blocked both the induction of SEFR during stress and its later expression, while artificial stimulation of pPVT in stress-naive mice was sufficient to recapitulate SEFR. Interestingly, pPVT inhibition or stimulation did not affect acquisition or expression of SEFL. In conclusion, our results indicate that sensitization of fear learning (SEFL) and sensitization of unlearned fear (SEFR) have distinct neural mechanisms. Our results identify pPVT hyperactivity as a mechanism for stress-induced sensitization of unlearned fear and highlight pPVT as a potential target for treating arousal and reactivity symptoms of trauma- and stressor-related disorders.

NeuID, a novel neuron-specific lncRNA, resolved a key epigenetic mechanisms linking gene silencing to Alzheimer's disease
The increasing evidence that non-coding RNAs can become deregulated during pathogenesis is dramatically expanding the space for drug discovery beyond the protein-coding genome. Long noncoding RNAs (lncRNAs) are emerging as key regulators of cellular function, yet most remain uncharacterized. Here, we identify a previously unstudied lncRNA, which we named Neuronal Identity (NeuID)--a conserved, brain-enriched transcript expressed exclusively in neurons. NeuID is downregulated in the brains of Alzheimers disease (AD) patients. Mechanistically, NeuID maintains neuronal identity by repressing developmental and glial genes via interaction with the PRC2 subunit EZH2 and regulation of H3K27me3. Knockdown of NeuID disrupts this repression, leading to impaired neuronal activity and memory formation. Importantly, CRISPRa-mediated NeuID overexpression restores neuronal function in A{beta}42-treated neurons. These findings identify NeuID as a critical regulator of neuronal plasticity and position it as a promising therapeutic target for AD.

One sentence summaryWe identify NeuID, a novel brain and neuron-specific long non-coding RNA downregulated in Alzheimers disease, as a key regulator of neuronal identity and a promising therapeutic target to restore neuronal function.

Resting-state functional dynamics alterations relate to plasma amyloid markers and explain memory impairments in the TgF344-AD model of Alzheimer's disease
Resting-state (RS) fMRI studies of Alzheimers diseases (AD) impact on brain function commonly use functional connectivity (FC), ignoring short-timescale network dynamics, captured by co-activation patterns (CAPs), shown to accurately classify transgenic rodents from the wild-type (WT). We acquired high temporal resolution RS-fMRI data in the TgF344-AD rat model at pre-plaque and plaque stages and delineated brain functional alterations using FC and CAPs. We also assessed plaque-stage blood amyloid levels and memory performance in the same animals and investigated the statistical relationship between pathological, RS-functional, and behavioral phenotypes. TgF344-AD (TG) rats had elevated blood amyloid levels, committed more working and reference memory errors and showed reduced hippocampal FC with the lateral cortical and default-mode-like network (DMLN) compared to WT at the plaque stage. They showed DMLN and hippocampal hyper- and hypo-activation at pre- and plaque stages respectively in multiple CAPs. While blood amyloid levels were explained better by plaque-stage, than pre-plaque stage, FC values and CAP activations, it was the pre-plaque stage, more than the plaque stage, CAP activations that accurately explained memory impairments. Our findings not only identify early signatures of AD in brain functional dynamics in this translational rat model but demonstrate their relevance for prognosis of memory deficits.

Human intention inference with a large language model can enhance brain-computer interface control: A proof-of-concept study
Brain-computer interface (BCI) control enables direct communication between the brain and external devices. However, BCI control accuracy with intention inferred from non-invasive modalities is limited, even when using data-driven approaches to tailor neural decoders. In this study, we propose a knowledge-driven framework for inferring human intention, leveraging large language models (LLMs) as an alternative to conventional data-driven approaches. We developed a neural decoder that integrates neural and oculomotor signals with contextual information using an LLM agent. Its feasibility was tested in a real-world BCI task to control a computer application. The LLM-based decoder achieved an average accuracy of 79% for responders (11 out of 20) in inferring the intention to select arbitrary posts in a social networking service. Ablation analyses revealed that integration of contextual information, multimodal signals, and empirical knowledge is critical for decoding accuracy. This study demonstrates the feasibility of a neural decoding framework using an LLM, paving the way for improved performance in BCI-driven external device operation for patients with disability.

HighlightsO_LILarge language models can infer human intent from neural and oculomotor signals integrated with screen context.
C_LIO_LIThe proposed model performed with higher accuracy compared with data-driven approaches.
C_LIO_LIAblation analyses revealed that integrating contextual information, multi-modal signals, and empirical knowledge is critical for decoding accuracy.
C_LI

eyeris: A flexible, extensible, and reproducible pupillometry preprocessing framework in R
Pupillometry provides a non-invasive window into the mind and brain, particularly as a psychophysiological readout of autonomic and cognitive processes like arousal, attention, stress, and emotional states. Pupillometry research lacks a robust, standardized framework for data preprocessing, whereas in functional magnetic resonance imaging and electroencephalography, researchers have converged on tools such as fMRIPrep, EEGLAB and MNE-Python; these tools are considered the gold standard in the field. Many established pupillometry preprocessing packages and workflows fall short of serving the goal of enhancing reproducibility, especially since most existing solutions lack designs based on Findability, Accessibility, Interoperability, and Reusability (FAIR) principles. To promote FAIR and open science practices for pupillometry research, we developed eyeris, a complete pupillometry preprocessing suite designed to be intuitive, modular, performant, and extensible (https://github.com/shawntz/eyeris). Out-of-the-box, eyeris provides a recommended preprocessing workflow and considers signal processing best practices for tonic and phasic pupillometry. Moreover, eyeris further enables open and reproducible science workflows, as well as quality control workflows by following a well-established file management schema and generating interactive output reports for both record keeping/sharing and quality assurance of preprocessed pupil data prior to formal analysis. Taken together, eyeris provides a robust allin-one transparent and adaptive solution for high-fidelity pupillometry preprocessing with the aim of further improving reproducibility in pupillometry research.

Ethanol inhibits dorsomedial striatum acetylcholine release
Alcohol use disorder (AUD) has severe adverse health and economic impacts totaling over $240 billion annually. Despite this, FDA approved treatments for AUD are limited. For this reason, further understanding of the neurobiological mechanisms in AUD is required for new treatments. An aspect of AUD involves deficits in behavioral flexibility that is similarly seen in animal models with depletion of dorsal striatal cholinergic interneurons (CINs). We found that acute EtOH (40 mM) inhibits the firing rate of dorsal striatal CINs, which are the primary source of acetylcholine (ACh) in the dorsal striatum. Additionally, we found through slice photometry recordings using an intensity-based ACh sensing fluorescent reporter (iAChSnFR) that acute EtOH (40 mM) inhibits dorsal striatal ACh release. In accord, in vivo fiber photometry with iAChSnFR also showed inhibition of ACh release following acute EtOH (2 g/kg ip). To induce EtOH dependence in mice, we used the chronic intermittent EtOH (CIE) vapor exposure model. Following CIE, we found that CIE-treated mice had a significant depression of ACh release compared to control mice in the dorsomedial but not dorsolateral striatum. Then, we performed stereological cell counts of CINs in CIE and control mice to examine the cause of this ACh deficit and found that CIE mice had a significant decrease in CINs in the dorsomedial but not dorsolateral striatum. In conclusion, our data show that EtOH inhibits dorsal striatal cholinergic signaling in a subregion specific manner that may contribute to AUD related behaviors.

Spatiotemporal Dynamics of Invariant Face Representations in the Human Brain
The human brain can effortlessly extract a familiar faces age, gender, and identity despite dramatic changes in appearance, such as head orientation, lighting, or expression. Yet, the spatiotemporal dynamics underlying this ability, and how they depend on task demands, remain unclear. Here, we used multivariate decoding of magnetoencephalography (MEG) responses and source localization to characterize the emergence of invariant face representations. Human participants viewed natural images of highly familiar celebrities that systematically varied in viewpoint, gender, and age, while performing a one-back task on the identity or the image. Time-resolved decoding revealed that identity information emerged rapidly and became increasingly invariant to viewpoint over time. We observed a temporal hierarchy: view-specific identity information appeared at 64 ms, followed by mirror-invariant representations at 75 ms and fully view-invariant identity at 89 ms. Identity-invariant age and gender information emerged around the same time as view-invariant identity. Task demands modulated only late-stage identity and gender representations, suggesting that early face processing is predominantly feedforward. Source localization at peak decoding times showed consistent involvement of the occipital face area (OFA) and fusiform face area (FFA), with stronger identity and age signals than gender. Our findings reveal the spatiotemporal dynamics by which the brain extracts view-invariant identity from familiar faces, suggest that age and gender are processed in parallel, and show that task demands modulate later processing stages. Together, these results offer new constraints on computational models of face perception.

Neural Traces of Forgotten Memories Persist in Humans and are Behaviorally Relevant
For a long time, forgetting has been taken as the dissipation of the neural memory traces (engrams). However, recent engram research in mice suggests that the engrams of forgotten memories do persist. This raises the question whether engrams underlying human episodic memories also persist despite forgetting? And do forgotten memories influence human behavior implicitly? To address this question, 40 men and women learned 96 face-object pairs. Using high-resolution functional magnetic resonance imaging at 7 Tesla we mapped the fate of the 96 memories at the systems level from learning to a 30-minute and onward to a 24-hour memory test. Upon each retrieval attempt, participants indicated whether they remembered or forgot the memory. Univariate and multivariate analyses of the functional brain data revealed that the engrams of forgotten memories remain implemented in the episodic memory network and continue to influence the accuracy of guessing responses at test. Overnight, the engrams of forgotten memories became implemented more deeply within bilateral hippocampus, while consciously accessible memories were neocorticalized overnight. The engrams of both consciously accessible and inaccessible (forgotten) memories were shifted from the 30-minute to the 24-hour retrieval within the right hippocampus and anterior cingulate gyrus such that pattern dissimilarities supported the correct retrieval responses at 24 hours. This is evidence that forgotten human episodic memories remain implemented in the episodic memory network and continue to influence human behavior implicitly.

Hierarchical diversification of instinctual behavior neurons by lineage, birth order, and sex
Brain regions devoted to instinctual behaviors, including the vertebrate hypothalamus and arthropod cerebrum, contain bespoke neural circuits dedicated to perceptual and internal regulation of many behavioral states. These circuits are usually complex in structure and contain an extensive diversity of cell types. The regulatory mechanisms that pattern circuits for instinctual behaviors have been challenging to elucidate. Here, we developed methods in Drosophila to transcriptionally profile identified neuronal stem cell lineages in the cerebrum. We applied this method to lineages that generate sex-differentiated neurons with known circuit roles. We identified 91 transcription factors that, in combinations of 6-8, delineate cerebral hemilineages - classes of postmitotic neurons born from the same stem cell and sharing Notch status. Hemilineages comprise the major anatomic classes in the cerebrum and these transcription factors are required to generate their gross features. We further identified 33 transcription factors characteristic of neuronal birth order within lineages; these subtly differentiate neuronal subtypes to provide common computational modules to circuits regulating different behaviors. Our findings suggest that hemilineage and birth order transcription factors operate in a hierarchical system to build, diversify, and sexually differentiate lineally-related neurons that compose complex instinctual circuits.

Human brain network control of creativity
The ability to think creatively is fundamental to human cognition, shaping how we interact with the world and navigate complex problems. Creativity depends on the interaction of multiple large-scale brain networks, but how these networks represent creative thought and distinguish it from other cognitive states remains unclear. Here, we use invasive intracranial recordings in the human brain to explore the network-level representations underlying creative thinking. Our results demonstrate that cortical networks in the human brain differentially encode distinct cognitive states: we found unique brain states underlying creative thinking vs. arithmetic calculations, particularly in the default mode network. Further, in the dorsal attention network, we uncovered nonlinear, high-dimensional representations of moment-by-moment fluctuations in creative performance. These representations define a neural creativity axis shared between network-level and single-neuron computations. Our findings reveal widespread neural representations of cognitive state and suggest distinct roles of specific cortical networks in controlling creativity, with the default mode network gating access to creative states and the dorsal attention network regulating the quality of creative output.

Accurate spatial localization of Allen Human Brain Atlas gene expression data for human neuroimaging
The Allen Human Brain Atlas has been a tremendously impactful resource in neuroimaging. The usefulness of this resource in neuroimaging arises from spatial coordinates of dissected tissue samples being provided in relation to a Montreal Neurological Institute (MNI)-space standard brain template, thereby allowing for the integration of gene expression and spatially standardized neuroimaging data. In fact, two previous sets of spatial coordinates exist, and surprisingly, the accuracy of these coordinates in placing dissected tissue samples in correct anatomical locations within MNI space have not been examined. Here, we show that there are significant inaccuracies in the two previous sets of coordinates, and provide a refined set of coordinates as a resource to the neuroscience community. We show (through analyses of meta-analytic data and a re-analysis of real study data) that using previous inaccurate coordinates can result in dramatically different genes being identified, which could compromise further downstream analyses.

Towards Unified Neural Decoding with Brain Functional Network Modeling
Recent achievements in implantable brain-computer interfaces (iBCIs) have demonstrated the potential to decode cognitive and motor behaviors with intracranial brain recordings; however, individual physiological and electrode implantation heterogeneities have constrained current approaches to neural decoding within single individuals, rendering interindividual neural decoding elusive. Here, we present Multi-individual Brain Region-Aggregated Network (MIBRAIN), a neural decoding framework that constructs a whole functional brain network model by integrating intracranial neurophysiological recordings across multiple individuals. MIBRAIN leverages self-supervised learning to derive generalized neural prototypes and supports group-level analysis of brain-region interactions and inter-subject neural synchrony. To validate our framework, we recorded stereoelectroencephalography (sEEG) signals from a cohort of individuals performing Mandarin syllable articulation. Both real-time online and offline decoding experiments demonstrated significant improvements in both audible and silent articulation decoding, enhanced decoding accuracy with increased multi-subject data integration, and effective generalization to unseen subjects. Furthermore, neural predictions for regions without direct electrode coverage were validated against authentic neural data. Overall, this framework paves the way for robust neural decoding across individuals and offers insights for practical clinical applications.

Dendritic computation for rule-based flexible categorization
A hallmark of intelligent behavior is the ability to flexibly respond to external sensory inputs based on dynamically changing rules. A central question is how neurons in the brain implement computations underlying intelligent behaviors. The neocortical pyramidal neurons use their elaborated dendritic arbors to segregate a plethora of inputs and dynamically integrate them--a process known as dendritic computation--which may play important roles in rule-dependent sensory processing. However, evidence directly linking dendritic computation with intelligent cognitive behaviors has been absent. Here we combine two-photon imaging and a rule-switching flexible categorization task in mice to show that a projectome-defined extratelencephalic (ET) cortical layer 5 (L5) neurons in the auditory cortex integrate dendritic rule information and somatic sensory input to enable rule-dependent flexible categorization. The apical dendrite and soma within the same ET neurons exhibit distinct compartmental representations for sensory and rule information, with the soma predominantly encoding sensory information and the dendrites representing inferred task rule. Simultaneous optogenetic dendritic inhibition and two-photon imaging revealed that dendritic rule coding is essential for somatic output of flexible categorization. Our findings indicate that nonlinear dendritic integration of rule and sensory information constitutes a neuronal computational mechanism underlying rule-switching flexible decision-making.

Two Sites, Two Languages: Dual-site tDCS and EEG Evidence for L1 Feature Transfer in L2
The neural bases of syntactic and semantic processing remain unclear. While prior transcranial direct current stimulation (tDCS) studies have targeted either the inferior frontal gyrus (LIFG) or left superior temporal gyrus (LSTG), we test whether dual stimulation of both alters second language (L2) anomaly based on native language (L1). Chinese and Korean participants evaluated Japanese sentence correctness; Japanese shares morpho-syntax with Korean and semantic radicals with Chinese. For the N400, dual anodal tDCS elicited a significant interaction stimulation and L1 (F (1,960.7) = 5.48, p = 0.0194), stimulation and sentence type (F (5,960.7) = 2.28, p = .0448), and L1 and sentence type (F (5, 960) = 2.28, p = 0.045). For the P600, a significant effect of stimulation and L1 interaction (F (1,1303.1) = 9.86, p = 0.0017), and stimulation x sentence type (F (5,1303.1) = 2.35, p = 0.039), suggesting L1 typology affected semantic integration during dual-stimulation tDCS.

Focused ultrasound neuromodulation of mediodorsal thalamus disrupts decision flexibility during reward learning
When learning to find the most beneficial course of action, the prefrontal cortex guides decisions by comparing estimates of the relative value of the options available. Basic neuroscience studies in animals support the view that the thalamus can regulate the activity within and across the prefrontal cortex. However, it is unknown if it can modify value-based decision making in humans. Understanding this is clinically important because disorders of value-based decision making underpin common motivational syndromes of apathy and impulsivity. Neuromodulation of the thalamus is routinely performed for motor symptoms and therefore could be a viable future target for these neuropsychiatric complications of neurodegenerative disease. Here, a group of patients (n=37) undergoing MR guided focused ultrasound for essential tremor, were tested using the restless bandit, a reward reinforcement learning task, immediately before and after thalamotomy. Performance was compared to a group of age- and gender-matched healthy controls (n = 32). Thalamotomy significantly impaired the proportion of switch choices during the task (P<0.001) without affecting overall performance. Importantly, this reduction in choice flexibility could not be attributed to diminished attention, as thalamotomy did not affect response time, accuracy, or lateralisation. A reinforcement learning model fitted to the patients' choices replicated the effect of thalamotomy when the model increased exploitation of the bandits' learnt value estimate. This shift in the explore-exploit trade-off, manifesting as reduced choice flexibility, co-varied with the extent of post-operative oedema extension into mediodorsal nucleus (R= 0.64, p<0.001), but no other, including the Ventral intermediate nucleus, targeted clinically during MRgFUS thalamotomy for tremor control. Furthermore, using a normative functional connectome, resting-state fMRI connectivity between the volume of oedema encroaching the mediodorsal nucleus and the prefrontal cortex, predicted individual patients change in bandit performance. Lastly, we used probabilistic tractography to confirm a thalamo-cortical circuit between the lateral portion of mediodorsal nucleus and the frontal polar cortex, which regulates decision flexibility and accounts for the thalamotomy induced behavioural effect observed in the restless bandit task. These findings confirm a causal role of the thalamus and specifically the mediodorsal nucleus, in regulating the extent to which value estimates are used to guide decisions and learning from reward. Neuromodulation of this target would be worthy of further investigation in patients with diminished motivation characteristic of apathy in conditions including Parkinson's disease.

Expectation effects based on newly learnt object-scene associations are modulated by spatial frequency
Objects typically appear within rich visual scenes. Some models of visual system function propose that scene information is extracted from low-spatial frequency components and rapidly propagates through the visual processing hierarchy. This contextual information may help bias perceptual inferences toward objects that are likely to appear within a scene, enacted via top-down feedback carrying predictions. We tested this hypothesised influence of low spatial frequency information through newly learnt predictive object-scene associations. We recorded electroencephalographic (EEG) data from 40 participants who viewed high-spatial frequency objects embedded in either low- or high-spatial frequency scenes. Object-scene pairings were probabilistically manipulated such that certain objects more frequently appeared in certain scenes. We trained classifiers on EEG data from object-only trials and tested them on object plus scene trials. We did not observe differences in classification accuracy across expected and unexpected objects for both low- and high-spatial frequency scenes, nor any interaction between spatial frequency and expectation. However, we observed expectation effects on event-related potentials. These effects arose at similar latencies for both low- and high-spatial frequency scenes. Together, we report evidence that expectations induced by object-scene pairings influence visually evoked responses but do not modulate object representations.