bioRxiv Subject Collection: Neuroscience's Journal

Latent encoding of movement in primary visual cortex
Neurons in the primary visual cortex (V1) are classically thought to encode spatial features of visual stimuli through simple population codes: each neuron exhibits a preferred orientation and preferred spatial frequency that are invariant to other aspects of the visual stimulus. Here, we show that this simple rule does not apply to the representation of major features of stimulus motion, including stimulus direction and temporal frequency (TF). We collected an extensive dataset of cat V1 responses to stimuli covarying in orientation, direction, spatial frequency, and TF to assess the extent of motion selectivity. In over half of V1, we found that the preferred direction changed with stimulus TF, revealing four distinct map motifs embedded within V1's functional architecture. Additionally, preferred TF was mostly uniform across the cortical surface. Despite the lack of spatial modulation for the preferred TF map and the lack of invariance for the preferred direction map, we found using convolutional neural networks that direction, TF and stimulus speed can be accurately decoded from V1 responses at all cortical locations. These findings suggest that subtle modulations of V1 activity may convey fine information about stimulus motion, pointing to a novel primary sensory encoding mechanism despite complex co-variation of responses to multiple attributes across V1 neurons.

SIMPL: Scalable and hassle-free optimization of neural representations from behaviour
High-dimensional neural activity in the brain is known to encode low-dimensional, time-evolving, behaviour-related variables. A fundamental goal of neural data analysis consists of identifying such variables and their mapping to neural activity. The canonical approach is to assume the latent variables are behaviour and visualize the subsequent tuning curves. However, significant mismatches between behaviour and the encoded variables may still exist --- the agent may be thinking of another location, or be uncertain of its own --- distorting the tuning curves and decreasing their interpretability. To address this issue a variety of methods have been proposed to learn this latent variable in an unsupervised manner; these techniques are typically expensive to train, come with many hyperparameters or scale poorly to large datasets complicating their adoption in practice. To solve these issues we propose SIMPL (Scalable Iterative Maximization of Population-coded Latents), an EM-style algorithm which iteratively optimizes latent variables and tuning curves. SIMPL is fast, scalable and exploits behaviour as an initial condition to further improve convergence and identifiability. We show SIMPL accurately recovers latent variables in biologically-inspired spatial and non-spatial tasks. When applied to a large rodent hippocampal dataset SIMPL efficiently finds a modified latent space with smaller, more numerous, and more uniformly-sized place fields than those based on behaviour, suggesting the brain may encode space with greater resolution than previously thought.

Nimodipine reduces microglial activation in vitro as evidenced by morphological phenotype, phagocytic activity and next generation RNA sequencing
Background: Nimodipine, an L-type voltage-gated calcium channel blocker, achieves vasorelaxation by suppressing Ca2+-dependent activation of cerebrovascular smooth muscle cells and is used to prevent delayed ischemic deficit following subarachnoid hemorrhage. Our preclinical drug repurposing studies raised the possibility that nimodipine may attenuate the pro-inflammatory shift in microglial function in response to brain injury. We analyzed the effects of nimodipine on activated microglia at the level of morphological and functional phenotypes, as well as their transcriptomic profile. Methods: Live brain slice preparations from C57BL/6 mice and primary microglia cultures from the cortex of neonatal Sprague Dawley rats were used. Brain slices were subjected to ischemia, and microglial cultures were activated with lipopolysaccharide (LPS; 20 ng/ml). Both preparations were treated with nimodipine (5-10-20 M). The degree of arborization was evaluated in Iba1-stained microglia and expressed as a transformation index (TI). Phagocytic activity of cultured microglia was visualized using fluorescent microbeads. TNF levels in the cultures were measured with ELISA. Total RNA was isolated from microglia and processed for next generation RNA sequencing to determine differentially expressed genes. Results: Nimodipine suppressed the ameboid morphological transformation and increased phagocytosis triggered by ischemia in brain slices and LPS in microglia cultures. At the transcriptional level, LPS resulted in a pro-inflammatory microglial phenotype, affecting the expression of cytokines, the complement system and phagocytosis-related genes. Focusing on the role of calcium in microglial activation, LPS increased RNA transcription of ionotropic purinergic and some TRP channels but decreased the expression of voltage- and ligand-gated calcium channels. In the endoplasmic reticulum, LPS downregulated gene expression of Ryr and IP3 receptors and increased transcription of the SERCA calcium pump gene. Nimodipine co-administered with LPS altered the expression of 110 genes in the opposite direction to LPS activation, of which at least 20 were associated with microglial immune response, 7 with cell adhesion and 2 with autophagy regulation. Conclusion: The effect of nimodipine goes beyond cerebral vasorelaxation. Nimodipine attenuates microglial activation by modulating Ca2+-dependent gene expression involved in intracellular signaling cascades to drive microglial immune responses. Consideration should be given to expanding the medical field of indication of nimodipine.

The organization of high-level visual cortex is aligned with visual rather than abstract linguistic information
Recent studies show that linguistic representations predict the response of high-level visual cortex to images, suggesting an alignment between visual and linguistic information. Here, using iEEG, we tested the hypothesis that such alignment is limited to textual descriptions of the visual content of the image and would not appear for their abstract textual descriptions. We generated two types of textual descriptions for images of famous people and places: visual-text, describing the visual content of the image, and abstract-text, based on their Wikipedia definitions, and extracted their relational-structure representations from a large language model. We used these linguistic representations, along with visual representation of the images based on deep neural network, to predict the iEEG responses to images. Neural relational-structures in high-level visual cortex were similarly predicted by visual-images and visual-text, but not abstract-text representations. These results demonstrate that visual-language alignment in high-level visual cortex is limited to visually grounded language.

Modeling neuron-astrocyte interactions in neural networks using distributed simulation
Astrocytes engage in local interactions with neurons, synapses, other glial cell types, and the vasculature through intricate cellular and molecular processes, playing an important role in brain information processing, plasticity, cognition, and behavior. This study aims to enhance computational modeling of local interactions between neurons and astrocytes to better understand how these interactions influence the activity regimes in neuron-astrocyte networks and overall brain function. We develop new algorithms for representing astrocytes and neuron-astrocyte interactions in distributed simulation code for large-scale spiking neuronal networks. These include an astrocyte model with calcium dynamics, an extension of a standard neuron model to receive calcium-dependent signals from astrocytes, connections that deliver these signals to neurons, and a connectivity concept that efficiently establishes interactions between groups of three elements (pre- and postsynaptic neurons and astrocytes) in large neuron-astrocyte populations. The latter extends the conventional concept of binary connectivity in neuronal circuits. We verify the efficiency of our reference implementation through a series of benchmarks that vary in computing resources and neuron-astrocyte network model sizes. Using the new technology, we reproduce experimental data on astrocytic effects on neuronal synchronization. Our in silico experiments demonstrate that astrocytes consistently induce local synchronization in groups of neurons across various connectivity schemes and global activity regimes. Changing only the efficiency of neuron-astrocyte interactions switches the global activity regime from asynchronous to network-wide synchronization. Our contribution supports reproducible and collaborative large-scale modeling projects of neuron-astrocyte circuits, addressing the growing interest in developing these models within the computational neuroscience community.

Human single-neuron activity is modulated by intracranial theta burst stimulation of the basolateral amygdala
The amygdala is a highly connected cluster of nuclei with input from multiple sensory modalities, particularly the ventral visual stream, and vast projections to distributed cortical and subcortical regions involved in autonomic regulation and cognition. Numerous studies have described the amygdala's capacity to facilitate the encoding of long-lasting emotional memories. Recently, direct electrical stimulation of the basolateral complex of the amygdala (BLA) in humans revealed a more generalized ability to enhance declarative memory irrespective of the emotional valence, likely by promoting synaptic plasticity-related processes underlying memory consolidation in the hippocampus and medial temporal lobe. These effects were achieved with rhythmic theta-burst stimulation (TBS), which is known to induce long-term potentiation (LTP), a key mechanism in memory formation. Emerging evidence suggests that intracranial TBS may also enhance memory specificity, evoke theta-frequency oscillations, and facilitate short-term plasticity in local field potential recordings. However, how amygdalar TBS modulates activity at the single-cell level and to what extent this modulation is associated with memory performance remains poorly understood. Here, we address this knowledge gap by conducting simultaneous microelectrode recordings from prefrontal and medial temporal structures during a memory task in which intracranial TBS was applied to the BLA. We observed a subset of neurons whose firing rate was modulated by TBS and exhibited highly heterogeneous responses with respect to onset latency, duration, and direction of effect. Notably, location and baseline activity predicted which neurons were most susceptible to modulation. These findings provide direct empirical support for stimulation-evoked modulation of single-neuron activity in humans, which has implications for the development and refinement of neuromodulatory therapies.

Functional segregation of cortical hand and speech areas by frequency detuning of an intrinsic motor rhythm.
Many decades after Penfield's (1937) classic depiction of the motor homunculus, it remains unclear how spatially contiguous and interconnected representations within human sensorimotor cortex might separate their activities to achieve the directed and precise control of distinct body regions evident in activities as different as typing and speaking. One long-standing but relatively neglected explanation draws from models of simple physical systems (like swinging pendulums) to posit that small differences in the oscillatory properties of neuronal populations (termed 'frequency detuning') can result in highly effective segregation of their activities and outputs. We tested this hypothesis by comparing the peak frequencies of beta-band (13-30 Hz) motor rhythms measured in a magnetoencephalographic neuroimaging study of finger and speech movements in a group of healthy adults and a group of typically developing children. Our results confirm a peak frequency task difference of about 1.5 Hz in the beta motor rhythms of both left and right hemispheres in adults. A comparable task difference was obtained in children for the left but not for the right hemisphere. These results provide novel support for the role of frequency detuning in the functional organisation of the brain and suggest that this mechanism should play a more prominent role in current models of bodily representations and their development within the sensorimotor cortex.

The Interaction Of Diet-Induced Obesity And Chronic Stress In A Mouse Model Of Menopause
Menopause is characterized by the cessation of ovarian hormone production. During postmenopause, cisgender women face increased risks of obesity, cognitive decline, and mood disorder. Mood disorders are associated with exposure to chronic stress. We investigated the combined effects of a high-fat diet (HFD) and chronic stress exposure in a mouse model of menopause using 4-vinylcyclohexene diepoxide (VCD), a selective ovotoxicant that gradually depletes ovarian follicles and hormones. Starting at 6 months, 82 female WT C57BL/6J mice received saline or VCD (130 mg/kg i.p.) 5 days per week for 3 weeks. One month after injection, mice were fed either low-fat diet (LFD) or HFD for 8 weeks followed by 6 weeks of chronic variable mild stress (CVMS). Post-CVMS, mice were either processed for gene expression of the anterodorsal BNST or behavior tests to assess cognitive and anxiety-related behaviors. Plasma samples were collected to analyze metabolic hormones and corticosterone levels. VCD-treated HFD-fed mice had higher fat and body mass, and elevated fasting glucose levels compared to controls and more pronounced avoidance behaviors and cognitive impairments. LFD-fed, VCD-treated mice exhibited less exploration of novel objects and open spaces compared to OIL and HFD counterparts. VCD elevated corticosterone levels on LFD and increased BNST Pacap gene expression on HFD. These findings highlight cognitive repercussions of estrogen deficiency and suggest a potential protective effect of a HFD against some of the adverse outcomes associated with menopause. Our study emphasizes the importance of considering dietary and hormonal interactions in the development of therapeutic strategies.

Melanin concentrating hormone projections to the nucleus accumbens enhance the reward value of food consumption and do not induce feeding or REM sleep
Regulation of food intake and energy balance is critical to survival. Hunger develops as a response to energy deficit and drives food-seeking and consumption. However, motivations to eat are varied in nature, and promoted by factors other than energy deficit. When dysregulated, non-homeostatic drives to consume can contribute to disorders of food intake, adding to the increasing prevalence of restrictive eating disorders and obesity. Melanin-concentrating hormone (MCH) neurons have been implicated in the regulation of feeding behavior, in addition to a number of other fundamental behaviors including sleep, anxiety, and maternal behavior. Several studies suggest that MCH peptide increases food consumption, while studies of MCH neurons show effects only on cued feeding, and others show no effect of MCH neuron manipulation on feeding. MCH neurons have widespread projections to diverse downstream brain regions yet few studies have investigated the function of specific projections or differentiated the behaviors they regulate. Here we use optogenetics, in combination with different behavioral paradigms, to elucidate the role of MCH projections to the nucleus accumbens (NAc) in sleep and feeding behavior. We show that MCH neurons projecting to the NAc do not induce changes in baseline feeding or REM sleep, but do enhance the preference for a food paired with optogenetic stimulation. Furthermore, this effect is diminished in female mice relative to males, in line with previous results suggesting sex differences in the functional role of MCH neurons. These results suggest that MCH projections to the NAc can enhance the rewarding value of consumed food.

Prefrontal cortex encodes behavior states decoupled from movement
Prefrontal cortex is often viewed as an extension of the motor system, but little is understood of how it relates to natural motor behavior. We therefore tracked the kinematics of freely moving rats performing minimally structured tasks and measured which aspects of behavior were read out in prefrontal neural populations. Naturalistic behaviors such as rearing or chasing a bait were each encoded by unique neural ensembles, but the behavioral representations were not anchored to posture or movement. Rather, the coding of kinematic features depended on their relevance to the animals' current behavior or which task the animal performed. Behavior-specific ensembles often preceded and outlasted physical actions and, accordingly, prefrontal population activity evolved at slower timescales than in motor cortex. These findings argue that prefrontal coding of behavior is not locked to motor output, and may instead reflect motivations to perform certain actions rather than the actions themselves.

ER Ca2+-levels control neuromodulator secretion by regulating STIM1 and L-type Ca2+-channel activity
Regulated secretion typically depends on activity-induced Ca2+ influx. However, in invertebrates, the endoplasmic reticulum (ER) plays a distinct role, particularly in the release of neuromodulators from dense-core vesicles (DCVs). Here, we investigated the role of the neuronal ER as a Ca2+ source for neuromodulator secretion in primary mouse neurons by directly monitoring ER and cytosolic Ca2+ dynamics, along with DCV exocytosis at single vesicle resolution. During neuronal activity, neurons with a low initial [Ca2+]ER took up Ca2+ into the ER, while those with a high initial [Ca2+]ER released ER Ca2+. These latter neurons showed more DCV exocytosis. Acute ER Ca2+ release by caffeine or thapsigargin application, resulted in minute increases in bulk cytosolic free Ca2+ that did not trigger DCV exocytosis. Remarkably, following ER Ca2+ depletion levels, activity-dependent Ca2+ influx and DCV exocytosis were reduced by 50-90%, while synaptic vesicle (SV) exocytosis was unaffected. L-type Ca2+-channel inhibition by nimodipine reduced DCV exocytosis and Ca2+ influx by 80-90 % without affecting SV exocytosis, a phenocopy of ER store depletion. In addition, introducing L-type channels lacking STIM1 interaction sites restored DCV fusion following ER store depletion. We conclude that the ER functions as a dynamic Ca2+ store serving both as a Ca2+ source or sink. Moreover, ER depletion activates a feedback loop that controls L-type Ca2+ channel activity, essential for DCV exocytosis.

Lysosomes cell autonomously regulate myeloid cell states and immune responses
Myeloid cells maintain tissue homeostasis via the recognition, engulfment, and lysosomal clearance of dying cells and cellular debris, which is often accompanied by changes from homeostatic to reactive states. While a role for phagocytic receptors in gating these transitions has been described, less is known about if and how lysosomes can contribute to transcriptional and functional plasticity. To determine how lysosomal health impacts myeloid cell states, we evaluated microglia and macrophages deficient for progranulin (encoded by Grn), a lysosomal protein with pleiotropic functions whose loss is associated with several neurodegenerative diseases. Single-cell RNA-sequencing of the aged mouse brain identified a Grn knockout (KO)-specific microglial subpopulation marked by high GPNMB expression that displays hallmarks of lysosomal dysfunction, including lipofuscin accumulation. Epigenetic analysis of aged microglia revealed MITF/TFE transcription factors as key mediators of the transcriptional states associated with Grn deficiency. In addition to identifying a core myeloid cell transcriptional response to diverse lysosomal stressors, targeted perturbations of various lysosomal properties in vitro uncovered a cell autonomous, TREM2-independent, response to lysosomal deacidification (via v-ATPase or VPS34 loss of function) that overlaps with Grn KO microglia phenotypes, including the induction of a lysosomal gene program, increased proliferation, and secretion of pro-inflammatory cytokines. Compound loss-of-function approaches established GPNMB upregulation upon lysosomal stress is required for the compensatory response to enhance lysosomal function via promoting acidification. Finally, pharmacological endolysosomal reacidification through sodium/proton exchanger inhibition partially rescued Grn KO microglia phenotypes. Overall, these data establish a fundamental link between lysosomal health and myeloid cell epigenetic, transcriptional, and functional states observed in neurodegeneration models.

Wireless recordings from dragonfly target detecting neurons during prey interception flight
Target interception is a complex sensorimotor behavior which requires fine tuning of the sensory system and its strategic coordination with the motor system. Despite various theories about how interception is achieved, its neural implementation remains unknown. We have previously shown that hunting dragonflies employ a balance of reactive and predictive control to intercept prey, using sophisticated model driven predictions to account for expected prey and self-motion. Here we explore the neural substrate of this interception system by investigating a well-known class of target-selective descending neurons (TSDNs). These cells have long been speculated to underlie interception steering but have never been studied in a behaving dragonfly. We combined detailed neuroanatomy, high-precision kinematics data and state-of-the-art neural telemetry to measure TSDN activity during flight. We found that TSDNs are exquisitely tuned to prey angular size and speed at ethological distances, and that they synapse directly onto neck and wing motoneurons in an unusual manner. However, we found that TSDNs were only weakly active during flight and are thus unlikely to provide the primary steering signal. Instead, they appear to drive the foveating head movements that stabilize prey on the eye before and likely throughout the interception flight. We suggest the TSDN population implements the reactive portion of the interception steering control system, coordinating head and wing movements to compensate for unexpected prey motion.

Chronic variable mild stress alters the transcriptome and signaling properties of the anterodorsal bed nuceleus of the stria terminalis in a sex-dependent manner
Chronic stress is a physiological state marked by dysregulation of the hypo-pituitary-adrenal axis and high circulating levels of stress hormones, such as corticosterone in mice or cortisol in humans. This dysregulated state may result in the development of mood disorders but the process by which this occurs is still unknown. The bed nucleus of the stria terminalis (BNST) serves as an integration center for stress signaling and is therefore likely an important area for the development of mood disorders. This project utilized a chronic variable mild stress (CVMS) paradigm to persistently stress mice for 6 weeks followed by RNA-Sequencing of the anterodorsal (ad) BNST and electrophysiology of corticotropin releasing hormone-expressing cells in the adBNST. Our results show significant sex-biases in the transcriptome of the adBNST as well as effects of CVMS on the transcriptome of the adBNST specifically in males. Female biased genes are related to synaptic transmission while male biased genes are related to RNA processing. Stress sensitive genes in males are related to synaptic transmission and synapse formation. Additionally, electrophysiology data showed that CVMS suppressed the M-current in males but not females. However, CVMS increased the strength of excitatory post-synaptic currents in females but not males. This suggests significant differences in how males and females process chronic stress. It also suggests that the BNST is more sensitive to chronic stress in males than in females.

Exploration of novel biomarkers for neurodegenerative diseases using proteomic analysis and ligand-binding assays
Neurodegenerative diseases are a major cause of morbidity and mortality worldwide, and their public health burden continues to increase. There is an urgent need to develop reliable and sensitive biomarkers to aid the timely diagnosis, disease progression monitoring, and therapeutic development for neurodegenerative disorders. Proteomic screening strategies, including antibody microarrays, are a powerful tool for biomarker discovery, but their findings should be confirmed using quantitative assays. The current study explored the feasibility of combining an exploratory proteomic strategy and confirmatory ligand-binding assays to screen for and validate biomarker candidates for neurodegenerative disorders. It analyzed cerebrospinal fluid (CSF) and plasma samples from patients with Alzheimer's disease, Parkinson's disease, and multiple sclerosis and healthy controls. The screening antibody microarray identified differentially expressed proteins between patients with neurodegenerative diseases and healthy controls. Quantitative ligand-binding assays confirmed that cluster of differentiation 14 (CD14) levels were elevated in CSF of patients with Alzheimer's disease, whereas osteopontin levels were increased in CSF of patients with Parkinson's disease. The current study demonstrated the utility of combining an exploratory proteomic approach and quantitative ligand-binding assays to identify biomarker candidates for neurodegenerative disorders. To further validate and expand these findings, large-scale analyses using well characterized samples should be conducted.

Indecision and recency-weighted evidence integration in non-clinical and clinical settings
Biases in information gathering are common in the general population and can reach pathological extremes in paralysing indecisiveness, as in obsessive-compulsive disorder (OCD). Here, we adopt a new perspective on information gathering and demonstrate an information integration bias where most recent information is over-weighted by means of evidence strength updates ({Delta}ES). In a large, crowd-sourced sample (N=5,237), we find reduced {Delta}ES-weighting drives indecisiveness along an OC spectrum. We replicate the attenuated {Delta}ES-weighting in a second lab-based study (N=105) that includes a transdiagnostic OC spectrum encompassing OCD and generalised anxiety patients. Using magnetoencephalography (MEG), we trace {Delta}ES signals to a late neural signal peaking around 920 ms. Critically, OC participants show an attenuated neural {Delta}ES signal in medio-frontal areas while other decision-relevant processes remain intact. Our findings establish biased information-weighting as a key driver of information gathering, where attenuated {Delta}ES can lead to indecisiveness across an OC spectrum.

Serum metabolome profiling in patients with mild cognitive impairment reveals sex differences in lipid metabolism
Alzheimers disease (AD) affects more women than men. Although women live longer than men, it is not longevity alone, but other factors, including metabolic changes, that contribute to the higher risk of AD in women. Metabolic pathways have been implicated in AD progression, but studies to date examined targeted pathways, leaving many metabolites unmeasured. Sex is often a neglected biological variable, and most metabolomic studies were not designed to investigate sex differences in metabolomic profiles. Here, we performed untargeted metabolomic profiling of sera from male and female patients with mild cognitive impairment (MCI), a common precursor to AD, and matched controls. We discovered significant metabolic changes in individuals with MCI, and found several pathways that were strongly associated with sex. Peptide energy metabolism demonstrated sexual dimorphism. Lipid pathways exhibited the strongest differences between female and male MCI patients, including specific phosphatidylcholine lipids, lysophospholipids, long-chain fatty acids, and monoacylglycerols. 1-palmitoleoyl glycerol and 1-arachidonoyl glycerol were higher in female MCI subjects than in male MCI subjects with no differences between control males and females. Conversely, specific dicarboxylic fatty acids were lower in female MCI subjects than male MCI subjects. In cultured astrocytes, 1-arachidonoyl glycerol promoted phosphorylation of the transcriptional regulator sphingosine kinase 2, which was inhibited by the transient receptor potential vanilloid 1 receptor antagonists, as well as chromatin remodelling. Overall, we identified novel sex-specific metabolites in MCI patients that could serve as biomarkers of MCI in both sexes, help further define AD etiology, and reveal new potential prevention strategies for AD.

Multimodal state-dependent connectivity analysis of arousal and autonomic centers in the brainstem and basal forebrain
Vigilance is a continuously altering state of cortical activation that influences cognition and behavior and is disrupted in multiple brain pathologies. Neuromodulatory nuclei in the brainstem and basal forebrain are implicated in arousal regulation and are key drivers of widespread neuronal activity and communication. However, it is unclear how their large-scale brain network architecture changes across dynamic variations in vigilance state (i.e., alertness and drowsiness). In this study, we leverage simultaneous EEG and 3T multi-echo functional magnetic resonance imaging (fMRI) to elucidate the vigilance-dependent connectivity of arousal regulation centers in the brainstem and basal forebrain. During states of low vigilance, most of the neuromodulatory nuclei investigated here exhibit a stronger global correlation pattern and greater connectivity to the thalamus, precuneus, and sensory and motor cortices. In a more alert state, the nuclei exhibit the strongest connectivity to the salience, default mode, and auditory networks. These vigilance-dependent correlation patterns persist even after applying multiple preprocessing strategies to reduce systemic vascular effects. To validate our findings, we analyze two large 3T and 7T fMRI datasets from the Human Connectome Project and demonstrate that the static and vigilance-dependent connectivity profiles of the arousal nuclei are reproducible across 3T multi-echo, 3T single-echo, and 7T single-echo fMRI modalities. Overall, this work provides novel insights into the role of neuromodulatory systems in vigilance-related brain activity.