bioRxiv Subject Collection: Neuroscience's Journal

Biologically Plausible Graph Neural Networks for Simulating Brain Dynamics and Inferring Connectivity
Understanding the complex connectivity of brain networks is essential for deciphering neural function and its disruptions in neurological disorders. Traditional graph theoretical approaches have provided valuable insights into the topological properties of neural networks, yet they often overlook the temporal dynamics intrinsic to neuronal activity. In this work, we introduce Cerebrum, a novel framework that seamlessly integrates biologically plausible Hodgkin-Huxley (HH) neuron models with Graph Neural Networks (GNNs) to simulate and infer synaptic connectivity in large-scale brain networks. Cerebrum leverages canonical network topologies: Erdos-Renyi, Small-World, and Scale-Free, to generate ground-truth connectivity matrices and employs advanced GNN architectures to accurately predict these connections from simulated neuronal activity patterns. By incorporating disease-specific parameter modifications, such as those mimicking Parkinson's disease and epilepsy, our framework enables the exploration of how pathological states influence network dynamics and connectivity inference. Additionally, we incorporate empirical synaptic data from C. elegans to enhance the biological fidelity of our simulations. Cerebrum is accompanied by an open-source computational toolkit, providing tools for neural dynamics simulation, connectivity inference, and interactive network perturbations. Our work represents a significant advancement in computational neuroscience, offering a biologically grounded machine learning framework for modeling and analyzing intricate brain network dynamics.

Effortful control and cortical brain structure in 5-year-old children: findings from the FinnBrain birth cohort study
The aim of this study was to explore the associations between an aspect of self-regulation (SR), effortful control (EC), and cortical brain structure in 5-year-old children. Efficient EC is a predictor of many attributes and important outcomes in life, such as social-emotional functioning, psychiatric and somatic health, finance, and criminal activity. The early brain correlates of EC are not widely studied, while a better understanding of them would aid in recognizing individuals at risk for maladaptive outcomes. Participants (N = 155) were a part of the FinnBrain Birth Cohort Study in Finland. T1-weighted brain magnetic resonance images were processed using FreeSurfer. The data was statistically analysed with a vertex-wise general linear model. At the age of 5 years, EC and its subscales, Attentional focusing, Inhibitory control (INH), Low intensity pleasure (LIP) and perceptual sensitivity, were assessed via parental report using The Children's Behaviour Questionnaire. We found positive associations between overall EC and cortical volume in the left superior parietal region and in the right inferior temporal region. We also found positive associations between EC and surface area on the left hemisphere in the superior parietal region. The findings were driven by the EC subscales of INH and LIP with INH linking positively with left surface area and volume, and LIP linking with left cortical volume. We extended the previous literature by shedding light on early structural brain correlates of EC in a large sample of typically developing 5-year-olds. The results differed significantly from previous findings in older children, highlighting the need for longitudinal studies to better understand the neural underpinnings of SR throughout development.

A molecularly defined brain circuit module for regulating the panic-like defensive state
Panic is an episode of strong defensive state, characterized by intense fear and severe physical symptoms such as elevated cardiorespiratory activities. How the brain generates panic state remains poorly understood. Here, we developed a robot-based experimental paradigm to evoke panic-like defensive state in mice. When stimulated by the robot, the mice exhibited jumping escapes and elevated cardiorespiratory activities. With this paradigm, we identified Cbln2-expressing (Cbln2+) neurons in the posterior hypothalamic nucleus (PHN) as a key neuronal population essential for the induction of panic-like defensive state. Activation of Cbln2+ PHN neurons induced behavioral and physical symptoms of panic-like defensive state. These neurons were strongly activated by noxious mechanical stimuli and encode jumping escape vigor. They were synaptically innervated by anxiety-associated brain areas and provoked panic-like defensive state via their projection to the periaqueductal gray. Together, our results reveal a molecularly defined circuit module that regulates the panic-like defensive state in mice.

Human shape perception spontaneously discovers the biological origin of novel, but natural, stimuli
Humans excel at categorizing objects by shape. This facility involves identifying shape features that objects have in common with other members of their class and relies--at least in part--on semantic/cognitive constructs. For example, plants sprout branches, fish grow fins, shoes are molded to our feet. Can humans parse shapes according to the processes that give shapes their key characteristics, even when such processes are hidden? To answer this, we investigated how humans perceive the shape of cells from the olfactory system of Xenopus laevis tadpoles. These objects are novel to most humans yet occur in nature and cluster into classes following their underlying biological function. We reconstructed 3D cell models through 3D-microscopy and photogrammetry, then conducted psychophysical experiments. Human participants performed two tasks: they arranged 3D-printed cell models by similarity and rated them along eight visual dimensions. Participants were highly consistent in their arrangements and ratings and spontaneously grouped stimuli to reflect the cell classes, unwittingly revealing the underlying processes shaping these forms. Our findings thus demonstrate that human perceptual organization mechanisms spontaneously parse the biological systematicities of never-before-seen, natural shapes. Integrating such human perceptual strategies into automated systems may enhance morphology-based analysis in biology and medicine.

Decreased BOLD signal variability in middle-aged and older adults on the Autism Spectrum
Purpose: Autism spectrum disorder (ASD) is a lifelong neurodevelopmental disorder. Preliminary evidence suggests an increased risk for accelerated or early-onset cognitive and neurological decline in ASD. While it is well established that brain development in children, adolescents and young adults with ASD diverges from neurotypical (NT) peers, it is unknown how brain function is impacted in older adults with ASD. Understanding age-related changes of brain function in ASD is crucial to establish best practices for cognitive and health screenings in adults with ASD and develop interventions that might reduce the risk of accelerated decline. Decreases in blood-oxygenation-level-dependent (BOLD) signal variability (BSV) in typical aging have been shown across multiple studies, likely reflecting declining Gamma-Aminobutyric Acid (GABA) activity, and is associated with poorer cognitive performance. We hypothesized that adults with ASD would show reduced BSV compared to the NT group, with steeper negative age associations in the ASD than NT group. Methods: The study assessed BSV in a cohort of adults (40-70 years), 28 with ASD and 39 age-matched NT. General linear models tested for main effects of diagnostic group (ASD, NT), age and group-by-age interactions (controlling for RMSD). Limitations: Our cross-sectional data and small sample size highlight the need for longitudinal analyses in larger cohorts, alongside exploring links to cognitive function. Additionally, psychotropic medications used by our cohort of adults on the autism spectrum may have affected BSV. Results: Significant group-by-age interactions were observed for the right insular, left temporal occipital fusiform, right frontal orbital and right inferior lateral occipital cortex, with BSV showing strong negative associations with age in the ASD but not NT group. Conclusion: These findings suggest that BSV decreases may occur earlier in adults on the autism spectrum compared to their neurotypical peers, possibly indicating accelerated aging. However, given limited prior research, additional longitudinal analyses will be necessary to determine if the results presented truly reflect accelerated aging or arise from lifelong persistent differences in brain function.

Development and early life stress sensitivity of the rat cortical microstructural similarity network
The rat offers a uniquely valuable animal model in neuroscience, but we currently lack an individual-level understanding of the in vivo rat brain network. Here, leveraging longitudinal measures of cortical magnetization transfer ratio (MTR) from in vivo neuroimaging between postnatal days 20 (weanling) and 290 (mid-adulthood), we design and implement a computational pipeline that captures the network of structural similarity (MIND, morphometric inverse divergence) between each of 53 distinct cortical areas. We first characterized the normative development of the network in a cohort of rats undergoing typical development (N=47), and then contrasted these findings with a cohort exposed to early life stress (ELS, N=40). MIND as a metric of cortical similarity and connectivity was validated by cortical cytoarchitectonics and axonal tract-tracing data. The normative rat MIND network had high between-study reliability and complex topological properties including a rich club. Similarity changed during post-natal and adolescent development, including a phase of fronto-hippocampal convergence, or increasing inter-areal similarity. An inverse process of increasing fronto-hippocampal dissimilarity was seen with post-adult aging. Exposure to ELS in the form of maternal separation appeared to accelerate the normative trajectory of brain development - highlighting embedding of stress in the dynamic rat brain network. Our work provides novel tools for systems-level study of the rat brain that can now be used to understand network-based underpinnings of complex lifespan behaviors and experimental manipulations that this model organism allows.

Protein kinase G inhibition preserves photoreceptor viability and function in a new mouse model for autosomal dominant retinitis pigmentosa
Retinitis Pigmentosa (RP) is the most common inherited retinal degeneration, characterized by an initial loss of rod photoreceptor cells. Photoreceptor cell death has been associated with high levels of cyclic guanosine-3', 5'- monophosphate (cGMP) in animal models of autosomal recessive RP (ARRP) and autosomal dominant RP (ADRP). cGMP analogues inhibiting protein kinase G (PKG) have been found to prevent rod degeneration in ARRP disease models, but their effects on ADRP are unknown. Here, we used the recently generated rhodopsin-mutant RhoI255d/+ ADRP mouse model to investigate cGMP-signaling and the effects of cGMP analogues targeting PKG. cGMP accumulation was investigated by retinal immunostaining in wild-type (WT), RhoI255d/+, and RhoI255d/I255d mice. The therapeutic efficacy of the cGMP analogues CN03 and CN238 was evaluated on organotypic retinal explant cultures derived from WT and RhoI255d/+ mice. Readouts included the TUNEL assay and immunostaining for cone arrestin-3. Downstream effectors of cell death were visualized using calpain, poly-ADP-ribose polymerase (PARP), and histone deacetylase (HDAC) in situ assays, as well as caspase-3 immunostaining. Photoreceptor function was assessed using micro-electroretinogram ({micro}ERG) recordings. When compared with WT, RhoI255d photoreceptors displayed cGMP accumulation in outer segments. In the RhoI255d/+ ADRP model, CN03 and CN238 significantly reduced the number of dying photoreceptors. However, the relatively small number of photoreceptors exhibiting caspase-3 activity was not changed by the treatment. Remarkably, CN238 effectively provided long-lasting neuroprotection of cone photoreceptors and preserved retinal light responsiveness of RhoI255d/+ retina. Overall, this study suggests caspase-independent but cGMP-dependent cell death as a dominant degenerative mechanism in the RhoI255d/+ ADRP mouse model. PKG inhibition demonstrated robust neuroprotection of both rod and cone photoreceptors, while the marked preservation of retinal function, especially with the compound CN238, highlighted cGMP analogues for the treatment of ADRP.

Frataxin deficiency in the astrocytes drives neurocognitive impairment in sickle cell disease mice
Individuals with sickle cell disease (SCD) suffer from a high burden of cerebrovascular lesions and cognitive impairment that vastly impact quality of life. Cerebrovascular lesions are characterized by microstructural neuroaxonal damage, but their pathogenesis has not been fully elucidated. Herein, we report that SCD mice (SS) have reduced expression of frataxin (FXN), a mitochondrial protein, in their astrocytes compared to control (AA) mice. Next, we generated chimeric mice with SS bone marrow and astrocyte-specific deletion of FXN (SSFXN-KO). Ex-vivo diffusion tensor magnetic resonance imaging and immunohistopathology of the brain showed that the SSFXN-KO mice have increased white matter neuroaxonal damage compared to the SS bone marrow chimera mice with wild-type FXN expression (SSFXN-WT). The SSFXN-KO mice also displayed poorer cognitive function as measured by the functional Y-maze and novel object recognition tests. Pharmacological induction of FXN by administration of insulin growth factor-1 improved cognitive function in the SSFXN-KO mice. Overall, our data demonstrate that FXN is a critical factor regulating neuroaxonal health and cognitive function in SCD mice. FXN may therefore be a novel pharmacologic target to prevent cerebrovascular complications in SCD.

Neuronal representation of the decisional reference point in monkeys
The reward reference point serves as a hidden benchmark for evaluating options in decision-making. Despite extensive behavioral evidence for reference-dependence, no neural representation of the reference point has been discovered. We analyzed single neuron activity from macaque monkeys performing a decision-making task designed to orthogonalize objective reward values from the reference point. Regression analyses of neuronal activity across six frontal brain regions identified a robust neural representation of the reference point in the ventral bank of anterior cingulate cortex (vbACC). Activity in the dorsal bank of anterior cingulate cortex and the dorsolateral prefrontal cortex, in contrast, encoded the reference-dependent subjective values of the rewards offered or obtained on each trial. The temporal dynamics of these signals and connections between these regions suggest a dedicated neural circuit implementing reference-dependent reward encoding, with the vbACC serving as the reference point signal source modulating activity in other frontal value-encoding areas.

Age-related Increase in Locus Coeruleus Activity and Connectivity with Prefrontal Cortex during Ambiguity Processing
Interpreting ambiguous environmental cues, like facial expressions, becomes increasingly challenging with age, especially as cognitive resources decline. Managing these challenges requires adaptive neural mechanisms that are essential for maintaining mental well-being. The locus coeruleus (LC), the brain's main norepinephrine source, regulates attention, arousal, and stress response. With extensive cortical connections, the LC supports adapting to cognitive demands and resolving conflicting cues from environment, particularly in later life. Previous research suggests that the LC interacts with the prefrontal cortex during high-conflict tasks. However, whether LC activity and its connectivity with the PFC support emotional ambiguity processing and contributes to emotional well-being in healthy aging remains unclear. To address this gap, we used 7T-MRI to examine LC function in 75 younger (25.8 +/- 4.02 years, 35 females) and 69 older adults (71.3 +/- 4.1 years, 35 females) during facial emotion-recognition task morphed with varying ambiguity: anchor (unambiguous happy or fearful), intermediate-ambiguity (30% happy-70% fearful and 40% happy-60% fearful expressions, in either direction), and absolute-ambiguity (50% happy-fearful). Behaviorally, participants had longer response times and lower confidence in the absolute-ambiguity condition, while older adults perceived ambiguous faces as happy more frequently than younger adults. Neuroimaging results revealed older adults exhibited greater LC activity and enhanced connectivity with dorsolateral PFC (dlPFC) during absolute-ambiguity compared to younger adults. This heightened connectivity in older adults was linked to better emotional resilience and mental well-being. These findings suggest greater LC activity supports managing cognitively demanding tasks, while enhanced LC-dlPFC connectivity helps maintain emotional well-being, underscoring the importance of this neural pathway for healthy aging.

Early-life Oxytocin Rescues Hippocampal Synaptic Plasticity and Episodic Memory in a Mouse Model of Fragile X Syndrome
Cognitive disabilities including impairments to episodic memory are debilitating features of autism spectrum disorder (ASD). Here we report that early-life treatment with oxytocin (OXT) fully restores episodic memory and associated synaptic plasticity in a rodent model of an ASD. Fmr1-knockout (KO) mice -- which mimic the single gene mutation in Fragile X Syndrome -- failed to encode three basic elements of an episode (identity, location, and temporal order) during a first time, unrewarded encounter with a set of cues. Intranasal administration of OXT during the second postnatal week eliminated each of these impairments in mice tested in adulthood. OXT treatment during the second, but not fifth, postnatal week also corrected pronounced defects in two distinct forms of hippocampal Long-Term Potentiation (LTP). Rescue of LTP in lateral perforant path (LPP) to dentate gyrus synapses was linked to recovery of NMDAR-gated synaptic responses, which are otherwise profoundly reduced in the mutant LPP. LTP induced by a threshold theta burst stimulation protocol in CA3-CA1 synapses was severely impaired in adult Fmr1-KOs as was a previously unreported post-induction growth phase for potentiation. Both effects were restored in adult Fmr1-KOs given early OXT treatment. Infusion of OXT into adult Fmr1-KO hippocampal slices normalized LTP in CA1 but had no effect on the defective potentiation (or NMDAR-mediated EPSCs) at LPP-dentate gyrus synapses. These results show that severe, autism-related defects in cognition critical to memory, as evident in a rodent model, are reversible and that an early-life therapeutic intervention can effect an enduring restoration of function.

Neural Mechanisms of tDCS: Insights from an In-Vivo Rodent Model with Realistic Electric Field Strengths
Introduction:Transcranial direct current stimulation (tDCS) is a non-invasive neuromodulation method using low amplitude current (1-2 mA) to create weak electric fields (<1 V/m) in the brain, influencing cognition, motor skills, and behavior. However, the neural mechanisms remain unclear, as prior studies used high electric field strengths (10-40 V/m) unrepresentative of human tDCS. Objective:This study aimed to develop an in-vivo rat model replicating human tDCS electric field strengths to examine effects of weak electric fields on cortical neurons. Method:Currents of 0.005-0.3 mA were applied in 9 rats, generating electric fields of 0.5-35 V/m in the somatosensory cortex. Neural activity across cortical layers was recorded using a multichannel silicone probe. Somatosensory evoked potentials (SSEP) elicited by foot shocks assessed membrane polarization. Regular spiking (RS) and fast-spiking (FS) neurons were identified via spike shapes. Effects of tDCS on SSEP, spontaneous spiking activity (SSA), and evoked spiking activity (ESA) were analyzed. Results:Anodal tDCS caused hyperpolarization (SSEP increase) in superficial layers and depolarization (SSEP decrease) in deeper layers, reversing asymmetrically for cathodal stimulation. Weak fields (<1 V/m) altered SSA in RS but not FS neurons, while stronger fields affected ESA in RS neurons. Effects correlated with field strength and were well described by linear mixed-effect models. Changes in SSA were correlated with changes in SSEP. Conclusion:This study demonstrates that realistic tDCS fields induce complex cortical polarization patterns linked to SSA changes. Increasing electric field strength amplifies effects, suggesting higher amplitude tDCS could enhance efficacy in humans.

Asymmetries in foveal vision
Visual perception is characterized by known asymmetries in the visual field; human's visual sensitivity is higher along the horizontal than the vertical meridian, and along the lower than the upper vertical meridian. These asymmetries decrease with decreasing eccentricity from the periphery to the center of gaze, suggesting that they may be absent in the 1-deg foveola, the retinal region used to explore scenes at high-resolution. Using high-precision eyetracking and gaze-contingent display, allowing for accurate control over the stimulated foveolar location despite the continuous eye motion at fixation, we investigated fine visual discrimination at different isoeccentric locations across the foveola and parafovea. Although the tested foveolar locations were only 0.3 deg away from the center of gaze, we show that, similar to more eccentric locations, humans are more sensitive to stimuli presented along the horizontal than the vertical meridian. Whereas the magnitude of this asymmetry is reduced in the foveola, the magnitude of the vertical meridian asymmetry is comparable but, interestingly, reversed: objects presented slightly above the center of gaze are more easily discerned than when presented at the same eccentricity below the center of gaze. Therefore, far from being uniform, as often assumed, foveolar vision is characterized by perceptual asymmetries. Further, these asymmetries differ not only in magnitude but also in direction compared to those present just ~4deg away from the center of gaze, resulting in overall different foveal and extrafoveal perceptual fields.

Disruptions in Primary Visual Cortex Physiology and Function in a Mouse Model of Timothy Syndrome
Timothy syndrome (TS) is a rare genetic disorder caused by mutations in the CACNA1C gene which encodes the L-type calcium channel -1 CaV1.2 subunit. While it is expressed throughout the body the most serious symptoms are cardiac and neurological. Classical TS1 and TS2 mutations cause prolonged action potentials (APs) in cardiomyocytes and in induced neurons derived from pluripotent stem cells taken from TS patients, but effects of TS mutations on neuronal function in vivo are not fully understood. TS is frequently associated with autistic traits, which in turn have been linked to altered sensory processing. Using the TS2-neo mouse model we analysed effects of the TS2 mutation on the visual system. We observed a widening of APs of pyramidal cells in ex vivo patch-clamp recordings and an increase in the density of parvalbumin positive (PV+) cells in the primary visual cortex. Neurons recorded extracellularly in vivo were less likely to respond to visual stimuli of low spatial frequency, but more likely to respond to visual stimuli of mid-to-high spatial frequency, compared to WT mice. These results point to a basic processing abnormality in the visual cortex of TS2-neo mice.

dFCExpert: Learning Dynamic Functional Connectivity Patterns with Modularity and State Experts
Characterizing brain dynamic functional connectivity (dFC) patterns from functional Magnetic Resonance Imaging (fMRI) data is of paramount importance in neuroscience and medicine. Recently, many graph neural network (GNN) models, combined with transformers or recurrent neural networks (RNNs), have shown great potential for modeling the dFC patterns. However, these methods face challenges in effectively characterizing the modularity organization of brain networks and capturing varying dFC state patterns. To address these limitations, we propose dFCExpert, a novel method designed to learn robust representations of dFC patterns in fMRI data with modularity experts and state experts. Specifically, the modularity experts optimize multiple experts to characterize the brain modularity organization during graph feature learning process by combining GNN and mixture of experts (MoE), with each expert focusing on brain nodes within the same functional network module. The state experts aggregate temporal dFC features into a set of distinctive connectivity states using a soft prototype clustering method, providing insight into how these states support different brain activities or are differentially affected by brain disorders. Experiments on two large-scale fMRI datasets demonstrate the superiority of our method over existing alternatives. The learned dFC representations not only show improved interpretability but also hold promise for enhancing clinical diagnosis.

Photoreceptor cell death mechanism in light damage
Light is one of the most ambient agents harmful to the retina. In Wistar rats prolonged exposure to 200 lux of LED light promotes photoreceptors cells death after 6 days. The possibility to know the cell photoreceptors death mechanism increases the likelihood to find a mixture of therapy with inhibitors or antagonists with anti-inflammatory and antioxidant effects. This study tested the cell death mechanism to elucidate the possible underlying molecular mechanisms. Adult male rats were exposed to 200 lux of LED light by different periods of constant light. The study focused on apoptosis and necroptosis pathways. Our findings reveal that necroptosis is an active pathway in retinal degeneration, occurring in both photoreceptors and glial cells, while apoptosis seem to be inactive during the studied time points.

Dynamics of efficient ensemble coding
Ensemble coding creates compressed representations of a stimulus array. When discriminating the ensemble average against a reference, however, items in the ensemble with feature values closer to the reference are typically weighed stronger. We have recently shown that this "robust averaging" behavior can be explained as a form of efficient coding, where the sensory encoding precision is dynamically adapted to efficiently represent ensemble stimuli according to their overall distribution relative to a trial-by-trial varying reference. However, the specific mechanisms underlying such dynamic, efficient ensemble coding remain unknown. Here, we demonstrate that the relative timing between the presentation of the reference and the stimulus ensemble strongly affects efficient ensemble coding. We systematically probed participants' decision behavior for varying time intervals between the presentation of the ensemble and the reference. We found that efficient ensemble coding was only clearly established when reference and ensemble were simultaneously presented. It was much weaker when the ensemble preceded the reference, and largely absent when the ensemble followed the reference. As captured by our model, reduced efficient ensemble coding thereby coincided with decreased decision accuracy in those asynchronous conditions. Our results indicate that any temporal offset between the ensemble and reference stimuli substantially disrupts the dynamic and efficient allocation of coding resource. This suggests that efficient ensemble coding is not the result of a preparatory attentional process nor due to evidence selection at the decision stage. Rather, it arises from a fast interaction between the simultaneously evoked, sensory representations of reference and ensemble stimuli.

In vivo changes in zebrafish anesthetic sensitivity in response to the loss of kif5Aa are associated with the alteration of mitochondrial motility
Anesthetic and sedative drugs are small compounds known to bind to hundreds of proteins. One intriguing binding partner of propofol is the kinesin motor domain, kif5A, a neuronal mitochondrial transport protein. Here, we used zebrafish WT and kif5Aa KO larval behavioral assays to assess anesthetic sensitivity and combined that with zebrafish primary neuronal cell culture to probe for alteration in mitochondrial motility. We found that the loss of kif5Aa increases behavioral sensitivity to propofol and etomidate, with etomidate hypersensitivity greater than propofol. In contrast, kif5Aa KO animals were resistant to the behavioral effects of dexmedetomidine. Finally, WT and kif5Aa KO larvae responded similarly to the behavioral effects of ketamine. Propofol inhibited the anterograde motility of mitochondria in WT zebrafish neurons, while etomidate inhibited mitochondrial motility in both anterograde and retrograde directions; neither drug altered mitochondrial motility in the kif5Aa knockout (KO) neurons. In contrast, dexmedetomidine enhanced retrograde mitochondrial motility in both WT and kif5Aa KO animals. Finally, ketamine had little significant effect on mitochondrial motility in either mutant or WT animals. These data demonstrate that each anesthetic/sedative drug affects the motor protein machinery uniquely and is associated with unique changes in behavior. Understanding how different anesthetic compounds alter neuron motor proteins will be important in defining how anesthetics alter neuronal signaling and energetic dynamics.

Dynamic Grouping of Ongoing Activity in V1 Hypercolumns
Neurons' spontaneous activity provides rich information about the brain. A single neuron's activity has close relationships with the local network. In order to understand such relationships, we studied the spontaneous activity of thousands of neurons in macaque V1 and V2 with two-photon calcium imaging. In V1, the ongoing activity was dominated by global fluctuations in which the activity of majority of neurons were correlated. Neurons' activity also relied on their relative locations within the local functional architectures, including ocular dominance, orientation, and color maps. Neurons with similar preferences dynamically grouped into co-activating ensembles and exhibited spatial patterns resembling the local functional maps. Different ensembles had different strengths and frequencies. This observation was consistent across all hypercolumn-sized V1 locations we examined. In V2, different imaging sites had different orientation and color features. However, the spontaneous activity in the sampled regions also correlated with the underlying functional architectures. These results indicate that functional architectures play an essential role in influencing neurons' spontaneous activity, and can be explained by a network model that integrates diverse horizontal connections among similar functional domains.

Energy efficiency and sensitivity benefits in a motion processing adaptive recurrent neural network
Motion processing is a key function for the survival of many organisms and is initially implemented in areas V1 and MT of the primate visual cortex. Advances in machine learning approaches have led to the development of motion processing neural networks that have elucidated several aspects of this process. However, it remains unclear how adaptation, a canonical function of sensory processing, influences motion processing. In this study, we developed two recurrent neural networks to study motion processing: MotionNet-R, a baseline model, and AdaptNet, a model that employs adaptive mechanisms inspired by biological systems. Both networks were trained on natural image sequences to estimate motion vectors. We found that both networks developed response properties that resembled those of neurons found in areas V1 and MT, e.g., speed tuning, and AdaptNet recapitulated the motion aftereffect phenomenon (i.e., the waterfall illusion). We show that the emergent computational properties that implement the phenomenon in AdaptNet confirm previous theoretical hypotheses. Further, we compared the performance of the two networks and found that AdaptNet processed motion more efficiently, operationalized as reduced activation. While AdaptNet incurred reduced accuracy in response to prolonged constant input, it was both more accurate and sensitive in response to changes in motion input. These results are consistent with theoretical explanations of adaptation as neural property that supports metabolic efficiency and increased sensitivity to change in the environment. Our findings provide novel insights into the neural mechanisms underlying motion adaptation and highlight the potential advantages of adaptive neural networks in modelling biological processes.

Spanning spatial scales with functional imaging in the human brain; initial experiences at 10.5 Tesla
One of the most important new frontiers in the effort to improve the spatial resolution and accuracy of imaging of human brain activity is the recent development of greater than 10 Tesla magnetic fields. Here we present initial results for 10.5 Tesla Blood Oxygenation Level Dependent (BOLD) based functional brain imaging (fMRI) of the human brain acquired with previously unavailable or difficult to attain spatial resolutions and functional contrast. We present data obtained with nominal isotropic resolutions ranging from 0.65 to 0.35 mm for partial brain coverage for stimulus evoked responses, and 0.75 mm for whole brain coverage to capture the spontaneous fluctuations that are the source of functional connectivity measures. The increasingly higher nominal resolutions (i.e. smaller voxel volumes and dimensions) employed in image acquisition were shown to correspond to real gains in resolution using image reconstruction methods developed to minimize blurring. Existence of supralinear gains in stimulus-evoked percent signal change and major improvements in statistical significance were evident relative to 7 Tesla, the most advanced commercially available ultrahigh magnetic field platform currently employed in human brain studies. These results were feasible due to gains in intrinsic signal-to-noise ratio, BOLD contrast, and image acceleration provided by the uniquely high magnetic field of 10.5 Tesla, and the use of novel high channel count arrays to capture these gains. The results provide a preview of the potential that will be available in the new era of greater than 10 Tesla human functional imaging, particularly for mesoscale functional organizations and connectivity.

The natural variability of a dexterous motor skill is stably encoded in the cortex of freely behaving mice
The nervous system enables precise, skilled movements across diverse contexts, consistently performed day after day. However, the extent to which the brain's neural activity patterns are specific to each instance of a movement or behavior, and how similar these patterns are across days, remains unclear. To address this, we record calcium fluorescence activity in the motor cortex of freely moving mice performing a self-initiated, skilled reach-to-grasp task. The high trial counts and single-trial variability in this task allow for a rigorous statistical analysis of moment-to-moment movement encoding across matched behavioral sets over five days. We found that single neurons in motor cortex encode the single trial (reach-to-reach) details of paw, digit, and head movements, and this encoding is stable over days. This suggests that the stable contribution of individual cells in the motor circuit underlies the generation of skilled movements, even in complex sensory-driven tasks like reach-to-grasp.

Mouse brain organoids model in vivo neurodevelopment and function and capture differences to human
In the last decade since their emergence, brain organoids have offered an increasingly popular and powerful model for the study of early development and disease in humans. These 3D stem cell-derived models exist in a newer space at the intersection of in vivo and 2D in vitro models. Functional benchmarking has so far remained largely uncharacterised however, leaving the extent to which these models may accurately portray in vivo processes still yet to be fully realised. Here we present a standardised unguided protocol to generate brain organoids from mice, the most commonly-used in vivo mammalian model; and in parallel establish a guided protocol for generating region-specific choroid plexus mouse organoids. Both unguided and guided mouse organoids progress through neurodevelopmental stages with an in vivo-like tempo and recapitulate species-specific characteristics of neural and choroid plexus development, respectively. Neuroepithelial cells generate neural progenitors that give rise to different neural subtypes including deep-layer neurons, upper-layer neurons, and glial cells. We further adapted protocols to prolong mouse cerebral organoid (CO) cultures as slices at the air-liquid interface (ALI), enhancing accessibility for long-term studies and functional investigations. In mature mouse ALI-COs, we observed mature glia, as well as synaptic structures and long-range axon tracts projecting to distant regions, suggesting an establishment and maturation of neural circuitry. Indeed, functional analyses with high-density multi-electrode arrays (HD-MEAs) indicate comparable activity to ex vivo organotypic mouse brain slices. Having established protocols for both region-specific and unpatterned mouse brain organoids, we demonstrate that their neurodevelopmental trajectories, and resultant mature features, closely mimic the in vivo models to which they are benchmarked across multiple biochemical, morphological, and functional read-outs. We propose that mouse brain organoids are a valuable model for functional studies, and provide insight into how closely brain organoids of other species, such as human, may recapitulate their own respective in vivo development.

A Vulnerable Subtype of Dopaminergic Neurons Drives Early Motor Deficits in Parkinson's Disease
In Parkinson's disease, dopaminergic neurons (DANs) in the midbrain gradually degenerate, with ventral substantia nigra pars compacta (SNc) DANs exhibiting greater vulnerability. However, it remains unclear whether specific molecular subtypes of ventral SNc DANs are more susceptible to degeneration in PD, and if they contribute to the early motor symptoms associated with the disease. We identified a subtype of Sox6+ DANs, Anxa1+, which are selectively lost earlier than other DANs, and with a time course that aligns with the development of motor symptoms in MitoPark mice. We generated a knock-in Cre mouse line for Anxa1+ DANs and showed differential anatomical inputs and outputs of this population. Further, we found that the inhibition of transmitter release in Anxa1+ neurons led to bradykinesia and tremor. This study uncovers the existence of a selectively vulnerable subtype of DANs that is sufficient to drive early motor symptoms in Parkinson's disease.

The impact of musical expertise on disentangled and contextual neural encoding of music revealed by generative music models
Music perception involves the intricate processing of individual notes and their contextual relationships within a piece. However, how the brain encodes and organizes these features, particularly in relation to musical expertise, remains unclear. Using noninvasive and invasive electrophysiological recordings alongside generative music models, we reveal that musicians exhibit neural encoding that is more attuned to the separation and integration of complex musical structures, with a pronounced left-hemispheric bias compared to non-musicians. Invasive recordings further highlight a hierarchical and spatially organized representation of musical features and context across brain regions. These findings advance our understanding of how the brain processes music and the role of musical training in shaping auditory cognition.

Conversion of silent synapses to AMPA receptor-mediated functional synapses in human cortical organoids
Despite the crucial role of synaptic connections and neural activity in the development and organization of cortical circuits, the mechanisms underlying the formation of functional synaptic connections in the developing human cerebral cortex remain unclear. We investigated the development of -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR)-mediated synaptic transmission using human cortical organoids (hCOs) derived from induced pluripotent stem cells. Two-photon Ca2 imaging revealed an increase in the frequency and amplitude of spontaneous activity in hCOs on day 80 compared to day 50. Additionally, spontaneous neural activity in late-stage hCOs, but not in early-stage hCOs, was blocked by N-methyl-D-aspartate receptor (NMDAR) and AMPAR antagonists. However, transsynaptic circuit tracing with G-deleted rabies viral vectors indicated a similar number of synaptic connections in early- and late-stage hCOs. Notably, chemical labeling demonstrated a significant increase in AMPAR expression on the postsynaptic membrane and colocalization with NMDAR in late-stage hCOs. These results suggest that hCOs progressively organize excitatory synaptic transmission, concurrent with the transition from silent synapses lacking AMPARs to functional synapses containing NMDARs and AMPARs. This in vitro model of human cortical circuits derived from induced pluripotent stem cells reflects the developmental programs underlying physiological transitions, providing valuable insights into human corticogenesis and neurodevelopmental disorders.

A modulatory attentional gate promotes recent memory expression in Drosophila
Using short-term memory (STM) requires that animals can distinguish memories of recent experience from those learned previously. In Drosophila the neuromodulator octopamine (OA) specifically affects STM through adrenergic-like receptors within neurons in the Mushroom Body (MB) network. However, why OA preferentially impacts STM remains unclear. A fly brain connectome reveals that two OA-VPM3 neurons provide most OA input to the MB network. Artificial OA-VPM3 activation during olfactory learning enhances STM and can divert odor salience. -adrenergic- like receptor knock-down in MB {beta} and {gamma} lobe MB Output Neurons (MBONs) generally impaired STM, while {beta}-adrenergic-like receptor loss from lobe MBONs enhanced memory. The identity and connectivity of the relevant MBONs suggests OA modulates an interaction between sites of STM and long-term memory (LTM) storage. Indeed, synthetic, odor, and learning-driven, OA-VPM3 activation temporarily blocks LTM expression. Therefore, OA reconfigures the MBON network so the fly prioritizes expression of recent over remote memories.

Aperiodic exponent of brain field potentials is dependent on the frequency range it is estimated
The aperiodic component of brain field potentials, like EEG, LFP and intracortical recordings, has shown to be a valuable tool in basic neuroscience and in clinical applications. Aperiodic activity is modeled as a power law of the power spectral density, with the aperiodic exponent as the key parameter. Part of the interest in this parameter lies in its proposed role as a marker of the balance between excitatory and inhibitory cortical activity. In theory, a perfect power law would imply that the same behaviour exists across all frequencies, however recent evidence has suggested that low and high frequency ranges could present different aperiodic exponents. To elucidate this, we systematically evaluated the relation between frequency range and aperiodic parameters using human resting-state intracortical recordings from 62 patients. We employed two distinct estimation methods, Specparam and IRASA. We found that aperiodic parameters were indeed dependent on frequency range. Specifically, we found that low frequency ranges displayed, on average, lower aperiodic exponents (flatter power spectral density) than high frequency ranges. This behaviour was consistent for Specparam and IRASA estimations in all frequency ranges compatible with EEG. Given that there is currently no consensus for a single frequency range to be used in either clinical or basic neuroscience, our results show that care should be taken when comparing aperiodic exponents derived from different frequency ranges. We believe our results also encourage further research into the possible roles that aperiodic exponents estimated from different frequency ranges could have in reflecting distinct aspects of cortical systems.

A biologically-inspired hierarchical convolutional energy model predicts V4 responses to natural videos
V4 is a key area within the visual processing hierarchy, and it represents features of intermediate complexity. However, no current computational model explains V4 responses under natural conditions. To address this, we developed a new hierarchical convolutional energy (HCE) model reflecting computations thought to occur in areas V1, V2, and V4, but which consists entirely of simple- and complex-like units like those found in V1. In contrast to prior models, the HCE model is trained end-to-end on neurophysiology data, without relying on pre-trained network features. We recorded 313 V4 neurons during full-color nature video stimulation and fit the HCE model to each neuron. The model's predicted optimal patterns (POPs) revealed complex spatiotemporal pattern selectivity in V4, supporting its role in representing space, time, and color. These findings indicate that area V4 is crucial for image segmentation and grouping operations that are essential for complex vision. Thus, responses of V4 neurons under naturalistic conditions can be explained by a hierarchical three-stage model where each stage consists entirely of units like those found in area V1.

Anxa1+ dopamine neuron vulnerability defines prodromal Parkinson's disease bradykinesia and procedural motor learning impairment
Progressive degeneration of dopamine neurons (DANs) defines Parkinson's disease (PD). However, the identity and function of the most vulnerable DAN populations in prodromal PD remain undefined. Here, we identify substantia nigra DANs with Annexin A1 (Anxa1) expression as selectively vulnerable across multiple prodromal PD models and significantly reduced in patient-derived DANs. We found that Anxa1+ DANs have a unique functional profile, as they do not signal reward or reinforce actions, and they are not necessary for motivated behavior. Instead, activity of Anxa1+ DAN axons correlates with vigorous movements during self-paced exploration, yet their silencing only disrupts a subset of action sequences that mirror a PD bradykinesia profile. Importantly, Anxa1+ DANs are essential for procedural learning in a maze task and for motor learning of dexterous actions. These findings establish the early vulnerability of Anxa1+ DANs in PD, whose function can explain prodromal bradykinesia and impairments in procedural motor learning.

Collapse of directed functional hierarchy under classical serotonergic psychedelics
It has been proposed that psychedelics induce profound functional changes to the hierarchical organisation of the human brain. Yet the term hierarchy is currently not well defined in neuroscience. Here, we use a precise definition of hierarchy, grounded in the theory of thermodynamics, which allows the quantification of temporal asymmetry in the directionality of information flow. We quantified the changes to the directed functional hierarchy of the brain under three classical serotonergic psychedelics - psilocybin, LSD and DMT. We found that all three psychedelics induce a reduction of the directed functional hierarchy, such that they display lower levels of global and network temporal asymmetry and a contraction of the two main patterns of variation of temporal asymmetry. Crucially, our results imply that the brain's directed functional hierarchy is collapsed under psychedelics, interpreted as yielding a more flexible brain. This enhanced flexibility in brain organisation under psychedelics may underpin the altered cognition and behaviour observed during these states, and, possibly afterwards, be suggestive of therapeutic potential.

Stress-induced mitochondrial fragmentation in endothelial cells disrupts blood-retinal barrier integrity causing neurodegeneration.
Increased vascular leakage and endothelial cell (EC) dysfunction are major features of neurodegenerative diseases. Here, we investigated the mechanisms leading to EC dysregulation and asked whether altered mitochondrial dynamics in ECs impinge on vascular barrier integrity and neurodegeneration. We show that ocular hypertension, a major risk factor to develop glaucoma, induced mitochondrial fragmentation in retinal capillary ECs accompanied by increased oxidative stress and ultrastructural defects. Analysis of EC mitochondrial components revealed overactivation of dynamin-related protein 1 (DRP1), a central regulator of mitochondrial fission, during glaucomatous damage. Pharmacological inhibition or EC-specific in vivo gene delivery of a dominant negative DRP1 mutant was sufficient to rescue mitochondrial volume, reduce vascular leakage, and increase expression of the tight junction claudin-5 (CLDN5). We further demonstrate that EC-targeted CLDN5 gene augmentation restored blood-retinal-barrier integrity, promoted neuronal survival, and improved light-evoked visual behaviors in glaucomatous mice. Our findings reveal that preserving mitochondrial homeostasis and EC function are valuable strategies to enhance neuroprotection and improve vision in glaucoma.