bioRxiv Subject Collection: Neuroscience's Journal

Developmental synaptic pruning in the olivo-cerebellar circuit sculpts predictive processing
Synaptic pruning is a dominant process of circuit assembly, known to eliminate a large proportion of synapses formed during development. Yet, very little is understood about how the drastic elimination of a large number of synapses affects circuit computation and behavior. Here, using the olivocerebellar system of larval zebrafish, we first establish the timelines of olivary climbing fiber to Purkinje neuron synaptic pruning. We then show that during this window, encoding of sensory-motor mismatches is improved, sensory representations are fine-tuned, and predictive signals are more robust. All of these lead to faster and better behavioral responses which are adaptable to current sensory input. Thus, developmental synaptic pruning makes circuits fit for complex computations such as predictive processing.

Auditory tract maturation in early infancy: Links to speech production
Infants exhibit remarkable language acquisition abilities, supported by highly plastic neural substrates that dynamically interact with early speech experiences. However, the developmental mechanisms of these neural substrates and their specific role in speech acquisition remain incompletely understood. Here, we present NeoAudi Tract (NAT), a robust automated toolbox for extracting the full set of auditory tracts in infants from birth to 24 months using $3$T diffusion MRI data. By characterizing the microstructural changes in these tracts, we demonstrate a gradual and continuous maturation process of the auditory system. Additionally, we identify significant correlations between auditory tract maturation and both textit{fine-motor skills} and textit{expressive language} $t$-scores from the Mullen Scales of Early Learning tests. Our findings highlight the role of the auditory system in speech production and indicate the intertwined development of auditory and motor systems that underlies speech acquisition, particularly during perceptual reorganization.

The Artists' Brain: A Data Fusion Approach to Characterize the Neural Bases of Professional Visual Artists
Artistic creativity relies on complex perceptual, cognitive and motoric functions, yet the specific neural characterization of being an artist remain incompletely understood. To fill this gap in the literature, the present study aims to characterize the gray matter (GM) and white matter (WM) contributions to professional visual artists as compared to non-artists controls. The MRI brain scans of 12 professional artists and 12 matched non-artists were analyzed via an unsupervised machine learning method known as Transposed Independent Vector Analysis (tIVA) to detect joint GM-WM networks. Two independent networks were found. The first network (IC2), more expressed in artists, included increased GM-WM concentration in regions associated with the Default Mode Network (DMN), Executive Control Network (ECN), and sensorimotor networks possibly related with augmented cognitive and ideational control, and increased perceptual-motor integration skills critical for creative tasks. The second network (IC8), less expressed in artists, included decreased GM-WM density in regions related to the Salience Network, such as the Anterior Cingulate Cortex, suggesting attentional regulation processes that may not be as central to visual artists. In sum, these results suggest that artists may rely on specialized brain networks, reflecting unique neural adaptations in individuals with pronounced creativity and extensive creative training.

VGluT3 BNST neurons transmit GABA and restrict feeding without affecting rewarding or aversive processing
The bed nucleus of the stria terminalis (BNST) is involved in feeding, reward, aversion, and anxiety-like behavior. We identify BNST neurons defined by the expression of vesicular glutamate transporter 3, VGluT3. VGluT3 neurons were localized to anteromedial BNST, were molecularly distinct from accumbal VGluT3 neurons, and co-express vesicular GABA transporter (VGaT). Cell-type specific presynaptic processes were identified in arcuate nucleus (ARC) and the paraventricular nucleus of the hypothalamus (PVN), regions critical for feeding and homeostatic regulation. Whole-cell patch-clamp electrophysiology revealed that, while these neurons co-express VGluT3 and VGaT, they functionally transmit GABA to both ARC and PVN, with rare glutamate co-transmission to ARC. Neuronal recordings of VGluT3 BNST neurons showed greater calcium-dependent signaling in response to sucrose consumption while sated compared with fasted. When fasted, optogenetic stimulation of BNST VGluT3 neurons decreased sucrose consumption using several stimulation conditions but not when stimulation occurred prior to sucrose access, suggesting that BNST VGluT3 activation concurrent with consumption in the fasted state reduces feeding. BNST VGluT3 activation during anxiety-like paradigms (novelty-suppressed feeding, open field, and elevated zero maze) and real-time place conditioning resulted in no changes in anxiety-like or reward/aversion behavior. We interpret these data such that VGluT3 BNST neurons represent a unique cellular population within the BNST that provides inhibitory input to hypothalamic regions to decrease feeding without affecting anxiety-like or reward/aversion behavior.

The Hypno-PC: Uncovering Sleep Dynamics through Principal Component Analysis and Hidden Markov Modeling of Electrophysiological Signals
The conventional approach to sleep analysis relies on pre-defined, visually scored stages derived from electrophysiological signals. This manual method demands substantial effort and is influenced by subjective assessments, implicitly assuming that these categories accurately reflect underlying biological processes. Recent advancements indicate that low-dimensional representations of complex brain activity can provide objective means of identifying brain states. These approaches can potentially uncover inherent patterns within sleep, offering valuable insights into its organization. In this study, we applied Principal Component Analysis (PCA) to spectral features extracted from high-density EEG, EOG, EMG, and ECG recorded overnight at both 30- and 4-second resolutions. Notably, the first principal component-the "Hypno-PC"-strongly aligns with the hypnogram at both time scales. Subsequently, we employed a Gaussian Hidden Markov Model (GHMM) to delineate discrete states in the PCA-transformed data and to quantify their temporal dynamics. Using minimal supervision (less than 0.5% of the data labeled) and a cross-subject approach, the model achieved alignment with standard sleep labels comparable to the typical inter-rater agreement. Finally, independent component analysis (ICA) was applied to the PCA space, decomposing it into an independent set of components that potentially represent distinct physiological processes. The integrated use of PCA, GHMM, and ICA provides a reproducible and scalable methodology that aligns with traditional sleep staging, while offering a more flexible and comprehensive perspective on sleep states. Our findings indicate that these data-driven, unsupervised methods effectively uncover the intrinsic dynamics of sleep, advancing automated sleep analysis and enhancing our understanding of sleep organization.

Frontal eye field neurons predict "anti-Bayesian" but not Bayesian judgments of visual stability across saccades
Bayesian models, in which priors are used to optimally compensate for sensory uncertainty, have had wide-ranging success in explaining behavior across sensorimotor contexts. We recently reported, however, that humans and monkeys use a combination of Bayesian and non-Bayesian strategies when making categorical judgments of visual stability across saccades. While they used priors to compensate for internal, movement-driven sensory uncertainty, consistent with Bayesian predictions, they decreased their use of priors when faced with external, visual image uncertainty, an "anti-Bayesian" adjustment consistent with the use of a simple classifier. Here, we tested for neural correlates of these Bayesian and classifier-based strategies in the frontal eye field (FEF), a prefrontal region shown to be important for the perception of visual stability across saccades. We recorded from single FEF neurons while two rhesus macaques performed the internal (motor) and external (image) noise tasks in each session, interleaved trial by trial. FEF activity correlated with and predicted the anti-Bayesian, but not the Bayesian, behavior. These results suggest that the two computational strategies for visual stability are implemented by distinct neural circuits and provide a first step toward an integrated understanding of the computational and neural mechanisms underlying visual perception across saccades.

CLN5 deficiency impairs glucose uptake in Batten disease
CLN5 disease, a form of juvenile dementia within the neuronal ceroid lipofuscinosis (NCL), is associated with mutations in the CLN5 gene encoding the lysosomal bis(monoacylglycero)phosphate (BMP) synthase, essential for BMP production and lysosomal function. Limited knowledge of cellular mechanisms and unclear drug targets hinder translating this to children's treatment, which remains symptomatic. We developed and characterized a new cln5 knock-out zebrafish model that replicates key features and molecular signatures of the human disease. Loss of Cln5 function in vivo altered axonal growth of retinal ON-bipolar cells revealing new disease features. Additionally, multi-omic analyzes at different developmental stages, revealed an impaired glucose metabolism as an original finding in NCL. This work demonstrates the profound metabolic impact of CLN5 dysfunction, offering a promising avenue toward targeted therapies for juvenile dementia.

Neural circuits between nodose ganglion and pulmonary neuroendocrine cells regulate lung inflammatory responses
The lungs are organs exposed to the external environment, and the air we inhale contains various pathogens, such as endotoxins. The vagus nerve, which innervates the lungs, may play a role in detecting pathogens that invade the lungs. Through transcriptome analysis, tissue clearance imaging, electrical excitability recording, and gene- and cell-specific knockout experiments, we found that vagus nerve endings innervate pulmonary neuroendocrine cells (PNECs). These nerve endings sense bacterial endotoxins via pain receptors (TRPA1) rather than toll-like receptors (TLR4), eliciting electrical excitation and enhancing the production of neuropeptides (CGRP) in the nodose ganglia. In turn, CGRP released by sensory neurons from the nodose ganglia promotes both neuropeptide production and the proliferation of PNECs, thereby amplifying endotoxin-induced lung inflammatory responses. This reveals that the neural circuits between the nodose ganglion and PNECs play a critical role in regulating lung inflammatory responses.

AA147 Alleviates Symptoms in a Mouse Model of Multiple Sclerosis by Reducing Oligodendrocyte Loss
Inflammation induced oligodendrocyte death and CNS demyelination, are key features of multiple sclerosis (MS). Inflammation-triggered endoplasmic reticulum (ER) stress and oxidative stress promote tissue damage in MS and in its preclinical animal model, experimental autoimmune encephalitis (EAE). Compound AA147 is a potent activator of the ATF6 signaling arm of the unfolded protein response (UPR) that can also induce antioxidant signaling through activation of the NRF2 pathway in neuronal cells. Previous work showed that AA147 protects multiple tissues against ischemia/reperfusion damage through ATF6 and/or NRF2 activation; however, its therapeutic potential in neuroinflammatory disorders remains unexplored. Here, we demonstrate that AA147 ameliorated the clinical symptoms of EAE and reduced ER stress, oligodendrocyte loss, and demyelination. Additionally, AA147 suppressed T cells in the CNS without altering the peripheral immune response. Importantly, AA147 significantly increased the expressions of Grp78, an ATF6 target gene, in oligodendrocytes, while enhancing levels of Grp78 as well as Ho-1, an NRF2 target gene, in microglia. In cultured oligodendrocytes, AA147 promoted nuclear translocation of ATF6, but not NRF2. Intriguingly, AA147 altered the microglia activation profile, possibly by triggering the NRF2 pathway. AA147 was not therapeutically beneficial during the acute EAE stage in mice lacking ATF6 in oligodendrocytes, indicating that protection primarily involves ATF6 activation in these cells. Overall, our results suggest AA147 as a potential therapeutic opportunity for MS by promoting oligodendrocyte survival and regulating microglia status through distinct mechanisms.

The role of STN beta oscillations on lower extremity muscle activity in Parkinsonian stepping
Freezing of gait (FOG) is a devastating symptom of Parkinson's disease (PD) often resulting in disabling falls and loss of independence. It affects half of patients, yet current therapeutic strategies are insufficient, and the underlying neural mechanisms remain poorly understood. This study investigated beta oscillation dynamics in the STN during different locomotor states, while examining the effects of levodopa. In particular, it aimed to identify pathological activity by analysing the relationship between the STN and lower limb muscles during stepping. Local field potentials (LFP) in the STN and muscle activity (EMG) of the gastrocnemius and peroneus longus were recorded in 14 PD patients during standing and stepping, ON and OFF levodopa. Levodopa reduced stepping variability, implying improved stepping abilities. Distinct STN beta patterns were observed between stepping and standing, with lower high-beta and higher low-beta during stepping compared to standing, suggesting a distinct role of these frequency bands in motor control during postural and movement states. Levodopa reduced low-beta but increased high-beta activity, highlighting a potential physiological function of high-beta in the STN during standing and stepping. In addition, step-phase specific effects of levodopa included reduced broad-beta band activity in the STN and lower limb muscles during the late-stance and pushing-off phase of the contralateral leg when ON medication. Further analyses suggest that pathological STN activity amplifies muscle activation around movement initiation, potentially reducing the ability of the patient to move freely. These findings offer insight for developing phase-specific stimulation strategies targeting STN beta oscillations during gait.

Assessing the role of BNST GABA neurons in backward conditioned suppression
Conditioned suppression is a useful paradigm for measuring learned avoidance. In most conditioned suppression studies, forward conditioning is used where a cue predicts an aversive stimulus. However, backward conditioning, in which an aversive stimulus predicts a cue, provides unique insights into learned avoidance due to its influence on both conditioned excitation and inhibition. We trained mice to consume sucrose in context A, associated an aversive stimulus in context B to few or many forward or backwards paired cues (CS+), and then tested for conditioned suppression in context A in response to the CS+. We found that few or many forward CS+ and few backward CS+ produced conditioned suppression, but many backwards cues did not. Administration of diazepam, a positive allosteric modulator of the GABA-A receptor, prevented conditioned suppression to the backward CS+ but not to the forward CS+. Furthermore, freezing behavior was observed in response to the forward CS+ but not the backward CS+, and diazepam had no effect on freezing or locomotion. We next examined BNST GABA neurons for potential sensitivity to backwards cues and conditioned suppression. VGaT BNST signaling increased in response to sucrose licks during the backward CS+ but not to licks outside the CS+ and not to the backward CS+ onset or offset. Using designer receptors, we found that BNST VGaT neuron activation, but not its inhibition, prevented backward conditioned suppression expression. We conclude that backward conditioned suppression is dependent on both positive allosteric modulation of GABA on GABA-A receptors by diazepam and BNST GABA neurons.

Perturbing whole-brain models of brain hierarchy: an application for depression following pharmacological treatment
Neural representation can extend beyond localised activity to encompass global patterns, where information is distributed across brain networks in a hierarchical manner. Recent research suggests that the hierarchy of causal influences shaping these patterns can serve as a signature of distinct brain states, with implications for neuropsychiatric disorders. Here, we first delve into how whole-brain models, guided by the Thermodynamics of Mind framework, can estimate the brain hierarchy of specific brain states, and how perturbations of such models can study the in-silico transitions to other states represented by static functional connectivity. We then show an application for major depressive disorder, where different brain hierarchical reconfigurations have been found following psilocybin and escitalopram treatments. We build whole-brain models of depressed patients before and after psilocybin and escitalopram interventions, and we carry a dynamic sensitivity analysis to explore the susceptibility of brain states and their drivability to healthier states. We show that susceptibility is on average reduced by escitalopram and increased by psilocybin, and that both treatments succeed in promoting healthier transitions. These results align with the post-treatment window of plasticity opened by serotonergic psychedelics, as well as with the similar clinical efficacy of both drugs observed in clinical trials.

The mechanosensory DEG/ENaC channel DEGT-1 is a proprioceptor of C. elegans foregut movement
The gastrointestinal tract is subjected to extensive mechanosensory stimulation during food ingestion. However, the identities of mechanosensory receptors in the enteric nervous system are largely unknown. The pharynx of C. elegans is a structurally and functionally homologous model of the vertebrate foregut, but is comprised of only 20 neurons that are embedded within the muscles and epithelial cells of the organ. Here we report that the DEG/ENaC family ion channel DEGT-1 is a proprioceptor of pharynx movement. DEGT-1 protein is expressed in 4 pharyngeal neurons (MI, M3, I4, and M5) and localized to their neuronal soma in direct contact with the collagenous pharyngeal basement membrane. degt-1 mutants display abnormally rapid feeding in the presence of food, causing global changes in lipid accumulation. degt-1 mutants also pump rapidly when pumping is induced by the presence of serotonin alone, suggesting that DEGT-1 is required for proprioception of pharyngeal pumping itself, rather than sensing ingested food. DEGT-1 is required in only two pharyngeal neurons (I4 and M5) to control pumping rate. Taken together, these results suggest that DEGT-1 modulates pharyngeal pumping rate by relaying proprioceptive feedback generated by the shear force of the pharynx against its own basement membrane. Thus, mechanosensors in enteric nervous systems modulate organ function not only by detecting forces from ingested contents, but also the movements of the organ itself.

Using green space virtual reality to prevent stress-induced working memory impairment
Exposure to green space is associated with both physical and mental health benefits, including the potential to buffer acute stress responses, positioning it as a promising non-pharmacological approach to protect cognitive functions against stress. However, urban residents often face significant barriers in accessing green spaces, which are not equitably distributed. Virtual Reality (VR) technology offers a potential solution by simulating green space that could be made accessible to a broader demographic. Here, we explored whether VR-stimulated green space could dampen biological stress responses (i.e., heart rate and cortical response) and prevent stress-induced working memory (WM) impairment. Healthy young participants underwent acute stress induction followed by 15 minutes of VR-based green space (N=36) or control empty space (N=30) intervention. Although we observed the expected stress-induced increase in heart rates and elevated cortisol levels under stress, VR green space exposure failed to temper cortisol responses compared to the control VR space. Further, VR green space exposure did not bring benefits to protect working memory performance under stress across three WM tasks. Applying a Bayesian analysis approach throughout enabled us to find substantial evidence for the absence of an effect of VR-based green space exposure on biological markers of acute stress responses and working memory performance. Our findings suggest that VR-generated green space may not effectively replicate the stress-buffering effects of actual green space exposure. We discussed the implications of our findings regarding the potential and limitations of using VR or green space exposure to buffer stress responses.

The Drosophila EGF domain protein Uninflatable sets the switch between wrapping glia growth and axon wrapping instructed by Notch
In the peripheral nervous system, sensory and motor axons are generally covered by wrapping glial cell processes. This neuron-glia interaction requires an intricate coordination of glial growth and differentiation. How this is controlled molecularly is largely unknown. At the example of Drosophila larval nerves, we show that glial growth is initially triggered by the FGF-receptor tyrosine kinase Heartless (Htl). In a screen for genes acting downstream of activated FGF-receptor, we identified the large membrane protein Uninflatable (Uif), which supports the formation of plasma membrane domains but not axon wrapping. Uif is also known to inhibit Notch. Surprisingly, we find that Notch signaling is required in postmitotic wrapping glia. While compromised Notch signaling results in a reduced wrapping efficiency, gain of Notch activity in wrapping glial cells leads to a hyperwrapping phenotype. Thus, Notch signaling is both necessary and sufficient for glial wrapping in Drosophila larvae. In addition, Notch suppresses both uif and htl function and thus stabilizes the switch between growth and differentiation. Given the general conservation of signaling mechanisms controlling glia development in mice and flies, similar mechanisms may act in the mammalian nervous system to control final glial differentiation.

Precise calcium-to-spike inference using biophysical generative models
The nonlinear intramolecular dynamics of fluorescent indicators of neural activity can distort the accurate estimate of action potential (spike) times. In order to develop a more accurate spike inference algorithm we characterized the kinetic responses to different calcium concentrations of three popular calcium indicators (GCaMP6f, jGCaMP7f, and jGCaMP8f) using in vitro stopped-flow and brain slice recordings. jGCaMP8f showed a use-dependent slowing of fluorescence responses that caused existing inference methods to generate numerous false positives. From these data we developed multistate models of GcaMPs and used them to create a Bayesian Sequential Monte Carlo (BiophysSMC) and machine learning (BiophysML) inference methods that largely reduced false positives and improved the spike time accuracy from publicly available ground-truth data. The BiophysSMC method detected individual spikes with an average uncertainty of 6 milliseconds, a 6-fold improvement over previous methods. Our framework highlights the advantages of physical model-based approaches over model-free algorithms.

Visual experience contributes to separation of face and language responses in the ventral stream
Human ventral occipitotemporal cortex (vOTC) contains specialized regions that support visual recognition of behaviorally relevant categories, including faces, written language and places. An open question is how experience interacts with innate constraints to enable functional specialization. We investigate this question by comparing vOTC function across sighted and congenitally blind adults. In sighted adults, a region in lateral vOTC called the fusiform face area (FFA) responds preferentially to faces, whereas distinct left-lateralized portions of vOTC respond to written language. In blind people, lateral vOTC responds to face touching, braille and speech, but their functional co-localization has not been tested. The same group of congenitally blind adults (n=20) touched faces and spatial layouts (Experiment 1) and performed a reading (braille) and spoken language task (Experiment 2). Sighted adults performed analogous tasks in the visual modality (n=28). Using within subject analyses, we replicate the separation of faces and written language in sighted adults: written language responses are found only in left vOTC and within that hemisphere they are separate from faces. By contrast, left and right vOTC responds to language in people born blind and in the left hemisphere face and language responses overlap. These findings suggest that visual experience contributes to segregating responses to face and language in vOTC. Co-localization of face and language responses suggests an innate predisposition for communication-relevant processing in lateral vOTC.