bioRxiv Subject Collection: Neuroscience's Journal
[Most Recent Entries]
[Calendar View]
Monday, November 25th, 2024
| Time |
Event |
| 3:21a |
Global Neuron Shape Reasoning with Point Affinity Transformers
Connectomics is a subfield of neuroscience that aims to map the brains intricate wiring diagram. Accurate neuron segmentation from microscopy volumes is essential for automating connectome reconstruction. However, current state-of-the-art algorithms use image-based convolutional neural networks that are limited to local neuron shape context. Thus, we introduce a new framework that reasons over global neuron shape with a novel point affinity transformer. Our framework embeds a (multi-)neuron point cloud into a fixed-length feature set from which we can decode any point pair affinities, enabling clustering neuron point clouds for automatic proofreading. We also show that the learned feature set can easily be mapped to a contrastive embedding space that enables neuron type classification using a simple KNN classifier. Our approach excels in two demanding connectomics tasks: proofreading segmentation errors and classifying neuron types. Evaluated on three benchmark datasets derived from state-of-the-art connectomes, our method outperforms point transformers, graph neural networks, and unsupervised clustering baselines. | | 3:21a |
Visual speech enhances auditory onset timing and envelope tracking through distinct mechanisms
Seeing the face of a speaker facilitates speech recognition in challenging listening environments. Prior work has shown that visual speech contains timing information to aid auditory speech processing, yet how these signals are integrated within the auditory system during audiovisual speech perception remains poorly understood. Observation of preparatory mouth movements may initiate phase reset of intrinsic oscillations, potentially sensitizing the auditory system for receptive speech processing, while observation of mouth movements post speech onset may facilitate entrainment to the speech envelope. Yet, little work has been done to test whether visual speech enhances encoding of auditory speech onset, speech envelope tracking, or both, and through independent or overlapping mechanisms. To investigate this, we examined the ways in which visual speech timing information alters theta band power and phase using human intracranial electroencephalography (iEEG) recordings in a large group of patients with epilepsy (n = 21). Prior to speech onset, preparatory mouth movements elicited theta phase reset (increased inter-trial phase coherence; ITPC) throughout the superior temporal gyrus (STG), which is thought to enhance speech onset encoding. Following speech onset, visual speech modulated theta ITPC only at anterior STG electrodes while theta power was modulated at posterior STG electrodes. Pre- and post-speech onset were spatially and temporally dissociated, consistent with the hypothesis that audiovisual speech onset encoding and envelope tracking mechanisms are partially distinct. Crucially, congruent and incongruent visual speech, designed here to have identical visual timing information about speech onset time, but different visual mouth evolution, produced only a small difference in the phase of theta band oscillations in the anterior STG, highlighting a more restricted role of visual speech in ongoing auditory entrainment. These results support the hypothesis that visual speech improves the precision of auditory speech encoding through two separate mechanisms, with auditory speech onset encoded throughout the entire STG and ongoing speech envelope tracking within anterior STG. | | 3:21a |
Separating three variability and noise sources in the response fluctuation of brain stimulation
Motor-evoked potentials (MEPs) are among the few directly observable responses to suprathreshold brain stimulation and serve a variety of applications. If the MEP size is graphed over the stimulation strength, they form an input-output (IO), recruitment, or dose-response curve. Previous statistical models with two variability sources inherently consider the small MEPs at the low plateau as part of the neural recruitment properties. However, recent studies demonstrated that small MEP responses are contaminated and over-shadowed by background noise of mostly technical quality and suggested that the recruitment curve should continue below this noise level. This work intends to separate physiological variability from background noise and improve the description of recruitment behaviour. We developed a model with three variability sources and a logarithmic logistic function without a lower plateau. Compared to previous models, we incorporated an additional source for background noise from amplifiers, electrode impedance, and remote bioelectric activity, which form the obesrved low-side plateau. Compared to the dual-variability source modes, our approach better described IO characteristics, evidenced by lower Bayesian Information Criterion scores across all subjects and pulse shapes. The model independently extracted hidden variability information across the stimulated neural system and isolated it from background noise, which led to an accurate estimation of the IO curve parameters. This new model offers a robust tool to analyse brain stimulation IO curves in clinical and experimental neuroscience, reducing the risk of spurious results from inappropriate statistical methods. By providing a more accurate representation of MEP responses and variability sources, this approach advances our understanding of cortical excitability and may improve the assessment of neuromodulation effects. | | 4:42a |
mTOR inhibition alleviated tau phosphorylation-induced mitochondrial impairment, oxidative stress, and cognitive impairment
Aim: Hyperphosphorylated tau plays a crucial role in the pathogenesis of Alzheimer's disease (AD). Whether mammalian target of rapamycin (mTOR) directly interacts with the Tau protein at Ser214, Ser356 and Thr231 is not clear. This study aimed to investigate whether mTOR-regulated tau phosphorylation disrupts mitochondrial dynamics and function and whether rapamycin, an mTOR inhibitor, can modulate tau phosphorylation levels and attenuate AD-related alterations. Methods: Adeno-associated virus (AAV) vectors were used to intracranially deliver the TauS214E/T231E/S356E (Tau3E) variant into 2-month-old C57BL/6 mice. The mice were intraperitoneally administered the mTOR inhibitor rapamycin for one week, followed by assessment via the Morris water maze test. Western blot analysis, immunofluorescence staining, and flow cytometry were employed to measure the expression levels of mTOR, p70S6K, and tau; mitochondrial dynamics; and reactive oxygen species (ROS) in HT22 cells and a mouse model overexpressing Tau3E, as well as in postmortem brain tissues from AD patients. Results: p-mTORS2448 colocalized with p-TauSer214, p-TauSer356, and p-TauThr231 in the hippocampal CA3 region of AD patients. HT22 cells and C57BL/6 mice overexpressing Tau3E presented elevated levels of p-mTOR, downstream target p-p70S6K, and ROS production; mitochondrial fragmentation; and p-TauSer214, p-TauSer356, and p-TauThr231. Rapamycin treatment partially mitigated the cognitive and molecular alterations in Tau3E mice. Conclusion: This study revealed a causal link between Tau phosphorylation at Ser214, Ser356, and p-Thr231 and mTOR upregulation and downstream impairments in ROS, mitochondrial dysfunction and cognitive function. Treatment using mTOR inhibitor rapamycin (i.p.) can alleviate impairment, reduce p-Tau and restore mitochondrial homeostasis, neuronal loss and cognitive impairment in mice. | | 4:42a |
Single-unit activations confer inductive biases for emergent circuit solutions to cognitive tasks
Trained recurrent neural networks (RNNs) have become the leading framework for modeling neural dynamics in the brain, owing to their capacity to mimic how population-level computations arise from interactions among many units with heterogeneous responses. RNN units are commonly modeled using various nonlinear activation functions, assuming these architectural differences do not affect emerging task solutions. Contrary to this view, we show that single-unit activation functions confer inductive biases that influence the geometry of neural population trajectories, single-unit selectivity, and fixed point configurations. Using a model distillation approach, we find that differences in neural representations and dynamics reflect qualitatively distinct circuit solutions to cognitive tasks emerging in RNNs with different activation functions, leading to disparate generalization behavior on out-of-distribution inputs. Our results show that seemingly minor architectural differences provide strong inductive biases for task solutions, raising a question about which RNN architectures better align with mechanisms of task execution in biological networks. | | 4:42a |
Controlling Reciprocity in Binary and Weighted Networks: A Novel Density-Conserving Approach
We introduce efficient Network Reciprocity Control (NRC) algorithms for steering the degree of asymmetry and reciprocity in binary and weighted networks while preserving fundamental network properties. Our methods maintain edge density in binary networks and cumulative edge weight in weighted graphs. We test these algorithms on synthetic benchmark networks-including random, small-world, and modular structures- as well as brain connectivity maps (connectomes) from various species. We demonstrate how adjusting the asymmetry-reciprocity balance under edge density and total weight constraints influences key network features, including spectral properties, degree distributions, community structure, clustering, and path lengths. Additionally, we present a case study on the computational implications of graded reciprocity by solving a memory task within the reservoir computing framework. Furthermore, we establish the scalability of the NRC algorithms by applying them to networks of increasing size. These approaches enable systematic investigation of the relationship between directional asymmetry and network topology, with potential applications in computational and network sciences, social network analysis, and other fields studying complex network systems where the directionality of connections is essential. | | 4:42a |
The Alzheimers Disease Risk Genes MS4A4A And MS4A6A Cooperate to Negatively Regulate Trem2 and Microglia states
Genetic variations in MS4A4A and MS4A6A are linked to the regulation of cerebrospinal fluid soluble TREM2 (sTREM2) levels and are associated with Alzheimer's disease (AD) risk and progression. Using CRISPR knockout and MS4A4A-degrading antibodies in primary human microglia, non-human primates (NHP), and a xenotransplantation model of amyloid pathology, we provide evidence that MS4A4A and MS4A6A are negative regulators of both the transmembrane and soluble TREM2 proteins. They also negatively regulate microglia proliferation, survival, metabolism, lysosomal function, energetics, phagocytosis, and disease-fighting states. Mechanistically, we find that MS4A4A exerts negative regulation by interacting with MS4A6A and protecting it from degradation. MS4A6A in turn forms a complex with and blocks the co-receptor DAP12, which is required for the stability, cell surface localization, and signaling of TREM2 and other receptors. Taken together, the data indicate that MS4A4A and MS4A6A are cooperating, post-transcriptional negative regulators of TREM2 and microglial function, and potential drug targets for AD. | | 4:42a |
Triple network dynamics and future alcohol consumption in adolescents
Background: Human neuroimaging increasingly suggests that the brain is best modeled as a highly interconnected and dynamic system. However, novel methodology for studying functional brain network dynamics have never been applied to the study of adolescent alcohol consumption. We sought to determine whether brain network dynamics are related to future drinking behavior in teenagers. Methods: Resting-state functional magnetic resonance imaging (fMRI) time series from 17-year-old non/low drinking participants (n=295) of the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) study were used to fit a Hidden Semi-Markov Model (HSMM). Regions of the default mode network (DMN), salience network (SN), and central executive network (CEN), collectively known as the Triple Network, were included in modeling. The HSMM identified the most-likely state sequence for each participant, a trajectory through distinct brain network states over the course of their fMRI scan. Poisson regression models were used to assess relationships between state sequence metrics and future drinking frequency. Potential sex differences in state sequence metrics or the relationship between sequence metrics and future drinking were assessed with permutation testing and interactions in regression models. Results: No sex differences in state sequence metrics were observed. However, the relationship between occupancy times and future drinking frequency differed by sex for two brain states. In the full sample, occupancy time in a state characterized by high interconnectivity between the SN and CEN was negatively associated with drinking. Occupancy time in a separate state characterized by high activation in the DMN and SN, but low activation in the CEN, was negatively associated with future drinking. Conclusions: Brain network dynamics may be useful neural markers of predisposition to drinking in adolescents. Brain states which make teens vulnerable or resilient to drinking may differ between sexes. | | 4:42a |
A meta-analysis of the effects of early life stress on the prefrontal cortex transcriptome suggests long-term effects on myelin
Background: Early life stress (ELS) refers to exposure to negative childhood experiences, such as neglect, disaster, and physical, mental, or emotional abuse. ELS can permanently alter the brain, leading to cognitive impairment, increased sensitivity to future stressors, and mental health risks. The prefrontal cortex (PFC) is a key brain region implicated in the effects of ELS. Methods: To better understand the effects of ELS on the PFC, we ran a meta-analysis of publicly available transcriptional profiling datasets. We identified five datasets (GSE89692, GSE116416, GSE14720, GSE153043, GSE124387) that characterized the long-term effects of multi-day postnatal ELS paradigms (maternal separation, limited nesting/bedding) in male and female laboratory rodents (rats, mice). The outcome variable was gene expression in the PFC later in adulthood as measured by microarray or RNA-Seq. To conduct the meta-analysis, preprocessed gene expression data were extracted from the Gemma database. Following quality control, the final sample size was n=89: n=42 controls & n=47 ELS: GSE116416 n=23 (no outliers); GSE116416 n=44 (2 outliers); GSE14720 n=7 (no outliers); GSE153043 n=9 (1 outlier), and GSE124387 n=6 (no outliers). Differential expression was calculated using the limma pipeline followed by an empirical Bayes correction. For each gene, a random effects meta-analysis model was then fit to the ELS vs. Control effect sizes (Log2 Fold Changes) from each study. Results: Our meta-analysis yielded stable estimates for 11,885 genes, identifying five genes with differential expression following ELS (false discovery rate< 0.05): transforming growth factor alpha (Tgfa), IQ motif containing GTPase activating protein 3 (Iqgap3), collagen, type XI, alpha 1 (Col11a1), claudin 11 (Cldn11) and myelin associated glycoprotein (Mag), all of which were downregulated. Broadly, gene sets associated with oligodendrocyte differentiation, myelination, and brain development were downregulated following ELS. In contrast, genes previously shown to be upregulated in Major Depressive Disorder patients were upregulated following ELS. Conclusion: These findings suggest that ELS during critical periods of development may produce long-term effects on the efficiency of transmission in the PFC and drive changes in gene expression similar to those underlying depression. | | 4:42a |
Brain Feature Maps Reveal Progressive Animal-Feature Representations in the Ventral Stream.
What are the fundamental units of representation in the primate visual brain? While objects have become an intuitive framework for studying neurons in many parts of cortex, it is possible that neurons follow a more expressive organizational principle, such as encoding generic features present across textures, places, and objects. In this study, we used multi-electrode arrays to record from neurons in early (V1/V2), middle (V4), and late (posterior inferotemporal cortex (PIT)) areas across the visual hierarchy, estimating the local operation of each neuron over the entire visual scene. These estimates, called "heatmaps," approximate the weight sharing operation of convolutional neural networks. We found that while populations of neurons across V1, V4, and PIT responded over the full scene, they focused on salient sub-regions within object outlines. The best captured object feature class belonged to animals, not general objects, as a trend that increased over the visual hierarchy. These results show that the monkey ventral stream is partially organized to encode local animal features over objects, even as early as primary visual cortex. | | 5:18p |
Dopamine alters motor learning performance in the presence and absence of feedback
Skilled motor performance is essential for survival. Indeed, we often not only choose to learn motor skills because of some external reward, but also because skilled movement, in and of itself, is satisfying. While dopamine is known to drive reward-based motor learning, it remains unclear whether dopamine is implicated in motor learning under conditions ostensibly driven by intrinsic rewards/motivation (i.e., in the absence of extrinsic feedback or reward). Here, we investigated the role of dopamine in motor skill learning guided by internally determined signals of performance success, using a task where learning occurred either in the absence or presence of feedback. We found that dopamine altered performance both in the presence and in the absence of information on task success. This provides direct causal evidence for a role of dopamine in motor learning driven by internal task goals. |
|