Browse Articles

Hypergravity hinders proper dendritic patterning of a sensory neuron in the nematode Caenorhabditis elegans

The contribution of audition and proprioception in unisensory and multisensory target reaching

Abstract Everyday actions often involve reaching for targets sensed by auditory and proprioceptive senses (reaching a ringing smartphone in the dark or tapping on it while holding it with the other hand). However, it is still unclear whether reaching performance toward auditory and proprioceptive targets is modality-specific and whether performance improves under multisensory compared to unisensory conditions. Here, we addressed these questions by measuring reaching performance toward auditory, proprioceptive, and combined audio-proprioceptive targets along the azimuth and depth dimensions. Accuracy was similar along the azimuth dimension, whereas precision was generally lower for auditory targets compared to proprioceptive and audio-proprioceptive targets. A second experiment investigated whether providing additional proprioceptive information while reaching for an auditory target could improve precision in subsequent auditory-only trials. A slight general improvement was observed, indicating that proprioceptive cues may help reduce spatial variability in auditory-guided actions, though not to the level seen in the multisensory condition. Overall, the results suggest that while the target modality slightly impacts movement accuracy, it has a significant impact on movement precision, with proprioceptive input playing a crucial role in enhancing precision. The concurrent availability of auditory and proprioceptive target information does not enhance precision beyond that achieved with proprioceptive information alone, whereas proprioception can modestly improve subsequent auditory-guided reaching.

Predictive expert assessments for large-scale battery storage system investments with conditional multi-facet fuzzy logarithmic least-squares and orthogonal metric robust aggregation

Design and lmplementation of wheel-leg-paddle lntegrated structure amphibious robot

Investigating the feasibility and safety of transcranial infraslow gray noise stimulation as a potential treatment for generalized anxiety disorder

Abstract Anxiety is a beneficial behavior that assists survival, although when high levels of anxiety persist it can hinder normal functioning. Despite numerous treatment options most individuals with Generalized Anxiety Disorder (GAD) continue to suffer from the disorder. A novel neuromodulatory treatment that safely and non-invasively alters pathological neural circuitry associated with GAD is required. Therefore, this study investigates the feasibility, safety, and effect of a new innovative High-Definition Transcranial Infraslow Gray Noise Stimulation (HD-tIGNS) targeting the anxiety network in people with GAD, in a delayed-start double-blinded randomized sham-controlled pilot trial ( N  = 22). HD-tIGNS was applied three times per week for three [delayed-start (following 3-weeks of actisham)], or six weeks (early-start). Anxiety questionnaires and electroencephalography were taken at baseline, mid-treatment, and post-treatment. On average, six participants were recruited per month with a mean treatment adherence of 99.5%, and an 8.3% dropout rate of enrolled participants. Clinically meaningful changes in GAD-7 were observed in 45% (early-start group) and 36% (delayed-start group) of participants following 6-weeks and 3-weeks of intervention respectively. No significant differences in brain activity or functional connectivity were observed. This study provides evidence supporting HD-tIGNS as a safe and feasible treatment approach for GAD, however future research should consider alternate network targets.

A novel hybrid metaheuristic algorithm and response surface methodology approach for predictive modeling and optimization of cementitious compressive strength

Research on an intelligent fan cultural sustainability design method based on CPO-CNN-LSSVM

Adaptive response to electrical pulse stimulation is impaired in FSHD myotubes by DUX4 gene network activation

Abstract Facioscapulohumeral dystrophy (FSHD) is one of the most common muscular dystrophies with no effective treatment. The disease is linked to abnormal derepression of DUX4 embryonic transcription factor in skeletal muscle, but at very low frequency. How this relates to the disease process and the exact nature of functional defects of FSHD muscle cells remain obscure. We used electrical pulse stimulation (EPS) substituting motor neuron activation to perform quantitative structural, functional, and gene expression analyses of in vitro differentiated control and FSHD skeletal myocytes. We found that EPS effectively stimulated muscle contractile gene expression and contractility activation in control cells, which are impaired in FSHD patient cells. This is accompanied by exacerbated organizational differences of sarcomeric striation. Importantly, this FSHD patient cell phenotype was confirmed in engineered mutant cells carrying D4Z4 repeat contraction and SMCHD1 mutation. Notably, FSHD patient/mutant cells fail to activate musclin/OSTN, an exercise-responsive muscle-protective myokine. Overexpression of LEUTX, a major downstream DUX4 target, was sufficient to recapitulate this phenotype, indicating that the phenotype is linked to DUX4 gene network activation. The results demonstrate DUX4-induced cell intrinsic functional defects of FSHD muscle cells in adaptive response to electrical stimulation, suggesting the possible activity-stimulated pathological mechanism of FSHD.

Multiple model trajectory poisson multi-bernoulli mixtures filter for tracking multiple maneuvering objects

Abstract Multi-object tracking (MOT) in cluttered and dynamic environments remains challenging, especially for maneuvering objects. While trajectory-based random finite set (RFS) filters provide principled solutions for trajectory estimation, they typically rely on a single motion model, limiting their adaptability. To address this gap, we propose the Multiple-Model Trajectory Poisson Multi-Bernoulli Mixture (MM-TPMBM) filter, which integrates jump Markov system (JMS) dynamics within the trajectory RFS framework. The filter enables closed-form Bayesian recursion for joint trajectory estimation and motion model switching. We derive its prediction and update equations, implement it using Gaussian mixtures for computational efficiency, and evaluate performance against benchmark filters (MM-PMBM, TPMBM, δ-GLMB) via Monte Carlo simulations. Results demonstrate that the MM-TPMBM filter achieves superior accuracy in trajectory estimation, localization, and cardinality, reducing the generalized optimal sub-pattern assignment (GOSPA) error by up to 24% and cardinality error by up to 52% compared to state-of-the-art methods, validating its robustness in complex tracking scenarios.

Fast and Site-Specific Covalent Targeting of Proteins by Arylfluorosulfate-Modified Aptamers

Spectral-spatial feature fusion for real-time facial expression recognition

Colored HOMFLYPT counts holomorphic curves

Given a link in the three-sphere, its Lagrangian conormal can be transplanted to the “resolved conifold,” which is a certain noncompact Calabi–Yau threefold. Here we show that, as predicted by Ooguri and Vafa using string theoretic arguments, the count of all holomorphic curves in the resolved conifold ending on this Lagrangian is, appropriately understood, the collection of the HOMFLYPT invariants of all colorings of the link. This generalizes our previous work, Skeins on branes, arXiv:1901.08027, which identified a curve count that captures the uncolored case. The main ingredient in the present work is a skein-valued multiple cover formula for an isolated embedded annulus.

Correction for Pospisil and Pillow, Revisiting the high-dimensional geometry of population responses in the visual cortex

Glomerular endothelial cells eliminate nicotinamide adenine dinucleotide to instruct CD103 ⁺ T cells in human lupus nephritis

Lupus nephritis (LN), which is characterized by the accumulation of DNA-containing immune complexes (ICs), is the leading cause of death in patients with systemic lupus erythematosus (SLE). While growing evidence highlights the central role of CD103 + T cells in shaping the immune landscape of regional tissues, mechanisms driving the cell differentiation in LN remain largely unexplored. In this study, we identified an increased frequency of CD4 + CD103 + T cells within the kidneys of SLE patients. Importantly, glomerular endothelial cells (ECs) from human LN tissues were found to promote the differentiation of CD4 + CD103 + T cells by upregulating B lymphocyte–induced maturation protein 1 (Blimp-1). Genetic knockdown of Blimp-1 in CD4 + T cells resulted in a reduced frequency of renal CD4 + CD103 + T cells and alleviated LN in humanized SLE chimeras. Mechanistically, LN-associated ECs, triggered by circulating DNA from SLE patients, exhibited elevated CD38 expression via the cGAS-STING signaling pathway. This facilitated the transfer of CD38 into CD4 + T cells through an exosome-dependent mechanism, leading to the depletion of nicotinamide adenine dinucleotide (NAD + ) levels in CD4 + T cells. The resulting NAD + depletion impaired the PARP1-mediated ADP-ribosylation of early growth response protein 1 (EGR1), which, in turn, enhanced Blimp-1 transcription and promoted CD103 + T cell differentiation. Targeting the cGAS/STING-CD38-EGR1 axis effectively reduced renal CD103 + T cell accumulation and inhibited LN progression in humanized SLE chimeras. Thus, ECs facilitate NAD + depletion to drive CD103 + T cell differentiation, presenting a cellular mechanism underlying LN pathogenesis and a potential therapeutic target for the clinical management of human LN.

Spiking world model with multicompartment neurons for model-based reinforcement learning

Brain-inspired spiking neural networks (SNNs) have garnered significant research attention in algorithm design and perception applications. However, their potential in the decision-making domain, particularly in model-based reinforcement learning, remains underexplored. In reinforcement learning, a world model refers to a predictive model that learns the environment’s dynamics and enables agents to simulate future trajectories in a latent space, thereby improving sample efficiency and long-horizon planning. The difficulty lies in the need for spiking neurons with long-term temporal memory capabilities, as well as network optimization that can integrate and learn information for accurate predictions. The dynamic dendritic information integration mechanism of biological neurons brings us valuable insights for addressing these challenges. In this study, we propose a multicompartment neuron model capable of nonlinearly integrating information from multiple dendritic sources to dynamically process long sequential inputs. Based on this model, we construct a spiking world model (Spiking-WM), which integrates a spiking state-space model, a spiking convolutional encoder, and a fully connected spiking network for policy learning, to enable model-based deep reinforcement learning with SNNs. We evaluated our model using the DeepMind Control Suite, demonstrating that Spiking-WM outperforms existing SNN-based models and achieves performance comparable to artificial neural network-based world models employing Gated Recurrent Units. Furthermore, we assess the long-term memory capabilities of the proposed model in speech datasets, including Spiking Heidelberg Digits dataset, Texas Instruments/Massachusetts Institute of Technology Acoustic-Phonetic Continuous Speech Corpus, and LibriSpeech 100h, showing that our multicompartment neuron model surpasses other SNN-based architectures in processing long sequences.

SPNS1 is an essential cellular factor for EV-A71 by acting as a transporter of viral pocket factor

Human enterovirus A71 (EV-A71) is a major cause of hand, foot and mouth disease. Cellular factors critical for EV-A71 infection remain enigmatic. Here, we performed CRISPR/Cas9 screens and identified sphingolipid transporter 1 (SPNS1) as an essential factor for EV-A71. SPNS1 deficiency inhibits infection of EV-A71 and 9 of 11 examined enteroviruses. Mechanistically, the endo/lysosomal localization of SPNS1 and the acidification of the endo/lysosomes are essential for SPNS1 to support EV-A71 infection. SPNS1 deficiency inhibits EV-A71 genomic RNA replication, but barely affects replication of EV-A71 RNA directly transfected into the cytoplasm. SPNS1 interacts with the EV-A71 capsid protein VP1 and entry receptor SCARB2 in the endo/lysosomes, where it acts as a transporter to release the viral pocket factor into the cytosol, leading to uncoating. Animal experiments show that SPNS1 deficiency results in reduced viral loads, pathological effects, and lethality following EV-A71 infection. Our findings collectively identified SPNS1 as a transporter of the EV-A71 viral pocket factor.

Computationally efficient whole-genome quantile regression at biobank scale

Genotype–phenotype associations can be context-dependent and dynamic in nature leading to heterogeneity of genetic effects across different parts of the phenotype distribution. Quantile regression, an alternative to linear regression for continuous phenotypes, is particularly well suited for detecting and characterizing heterogeneous genotype–phenotype associations. Here, we propose a computationally efficient whole-genome quantile regression technique, Regenie.QRS, for biobank-scale genome-wide association studies (GWAS) data with genetic structure. Our approach first estimates the polygenic effect, and then incorporates this effect as an offset in the nonmixed quantile regression model. Our simulations demonstrate robust control of type I error and higher power to detect heterogeneous associations relative to linear regression in GWAS and improved power over the marginal quantile regression tests. We present applications using data from the UK Biobank and the ProgeNIA/SardiNIA project, where we show the advantages of Regenie.QRS in identifying and characterizing heterogeneous genetic effects. To cite just one interesting example, using quantile regression, we are able to show that even though variants at the G6PC2 locus increase glucose levels, their effects are much stronger at lower quantiles of glucose level distribution than at higher quantiles, suggesting that G6PC2 may serve as a guardian against low glucose levels without driving dangerous hyperglycemia, which may explain the previously reported lack of association with diabetes risk. Beyond human genetics, our approach can be applied to plant and animal genetic studies to improve selective breeding and conservation efforts.

Evaluating large language models in biomedical data science challenges through a classroom experiment

Large language models (LLMs) have shown remarkable capabilities in algorithm design, but their effectiveness in solving data science challenges in real-world settings remains poorly understood. We conducted a classroom experiment in which graduate students used LLMs to solve biomedical data science challenges on Kaggle, focusing on tabular data prediction. While their submissions did not top the leaderboards, their prediction scores were often close to those of leading human participants. LLMs frequently recommended gradient boosting methods, which were associated with better performance. Among prompting strategies, self-refinement, where the LLM improves its own initial solution, was the most effective, a result validated using additional LLMs. While LLMs are capable of handling more complex data science tasks beyond tabular data prediction, their performance is substantially worse. These findings demonstrate that LLMs have the potential to design competitive machine learning solutions, even when used by nonexperts.

Building hierarchically nested structure by rapid neural sequences

Hierarchically nested structures are fundamental to human cognition, enabling complex behaviors across domains including language, planning, and mathematics. However, the neural mechanisms that enable the flexible construction of these hierarchical structures are poorly understood. Here, we designed a task where participants mentally built sequences with nested, multidepth structures by recursively applying a fixed set of rules. Using magnetoencephalography, we find that the brain constructs nested hierarchies through rapid neural sequences that perform two recurring generative operations. The first operation identifies the hierarchy depth of a symbol and is associated with increased ripple-band power; while the second arranges the symbol into its correct order at that level, a process that scales with the number of depths, also positively correlated with planning time. These results reveal a fundamental neural computation for transforming sensory information into structured representations, which is essential for higher-order cognition.

TAPT1 interacts with SUCO to maintain the homeostasis of newly synthesized proteins and brain development in mice

Genetic mutations in Tapt1 cause complex skeletal dysplasia and structural brain abnormalities. Although the pathogenesis underlying skeletal dysplasia has been explored, the functions and potential mechanisms of transmembrane anterior–posterior transition 1 (TAPT1) during brain development have not been reported. Here, we show that the brains of Tapt1 conditional knockout mice exhibit severe neurodevelopmental defects, including impaired proliferation and differentiation of neural progenitor cells and defects in dendritic and synaptic development, leading to severe microcephaly, motor dysfunction, and early death. Mechanically, we reveal that TAPT1 interacts with SUCO in the endoplasmic reticulum to maintain newly synthesized proteins, including those important for brain development. The TAPT1–SUCO complex plays an essential role in the homeostasis of newly synthesized proteins, and its loss causes overactivated protein degradation, as well as impaired endoplasmic reticulum-to-Golgi trafficking and organelle structures. Our results thus provide insights into the pathogenesis of TAPT1 and SUCO mutation–associated diseases that share similar pathologies.