Science Journals

Apparent neural encoding of future words may arise from the statistical structure of language itself, rather than from predictive computations in the brain.

Voltage-gated K⁺ channels play central roles in human physiology, in health, and disease. A repertoire of inhibitors that are both potent and specific would, therefore, be of great value. RY785 has been described as promising in this regard, as it selectively inhibits channels in the Kv2 subfamily with high potency. Its mechanism of action has not yet been determined at the molecular level, but functional studies indicate it differs from those of less specific inhibitors, such as quaternary-ammonium compounds or aminopyridines. To examine this mechanism at the single-molecule level, we have carried out a series of all-atom molecular dynamics simulations based on the structure of the Kv2.1 channel in the ion-conducting state. The simulations demonstrate both RY785 and tetraethylammonium spontaneously enter the channel interior through the cytoplasmic gate, but with distinct effects. Tetraethylammonium binds to a site adjacent to the selectivity filter, on the pore axis, thus blocking the flow of K⁺ ions. RY785, by contrast, binds to the channel walls, off-axis, and allows K⁺ flow while the gate remains open. This observation indicates RY785 inhibits Kv2.1 by fostering the occlusion of the gate, through a network of hydrophobic interactions therein, explaining why it also modulates the voltage-sensing mechanism of the channel, 3 nanometers away.

The qualities of antibody (Ab) responses provided by B lymphocytes and their plasma cell (PC) descendants are crucial facets of responses to vaccines and microbes. Metabolic processes and products regulate aspects of B cell proliferation and differentiation into germinal center (GC) and PC states along with Ab diversification. However, there is little information about lymphoid-cell-intrinsic functions of enzymes that mediate ether lipid biosynthesis. Imaging mass spectrometry (IMS) results had indicated that concentrations of a number of these phospholipids were substantially enhanced in GC compared to the background average in spleens, but it was unclear if biosynthesis in B cells was a basis for this finding, or whether cell-intrinsic biosynthesis contributes to B cell physiology or Ab responses. Ether lipid biosynthesis can involve the enzyme PexRAP, encoded by the <i>Dhrs7b</i> gene. Using IMS and immunization experiments in mouse models with inducible <i>Dhrs7b</i> loss of function, we now show that B-lineage-intrinsic expression of PexRAP promotes the magnitude and affinity maturation of a serological response. Moreover, the data revealed a <i>Dhrs7b</i>-dependent increase in ether phospholipids in primary follicles with a more prominent increase in GC. Mechanistically, PexRAP impacted B cell proliferation via enhanced survival associated with controlling levels of ROS and membrane peroxidation. These findings reveal a vital role of this peroxisomal enzyme in B cell homeostasis and the physiology of humoral immunity.

During mouse development, hematopoietic stem and progenitor cells (HSPCs) originate from hemogenic endothelial cells (ECs) through a process of endothelial-to-hematopoietic transition. These HSPCs are thought to fully sustain adult hematopoiesis. However, it remains unknown whether adult ECs retain hemogenic potential. Here, we used in vivo genetic lineage tracking at population and single-cell (sc) levels, scRNA sequencing, and bone marrow (BM) transplantation to detect hemogenic ECs in adult mice. We identify and characterize BM-resident, adult <i>Cdh5</i>/VE-Cadherin⁺ ECs that produce hematopoietic cell-progeny in vitro and in mice. These adult hemogenic ECs and their hematopoietic cell progeny give rise to hematopoietic cells following adoptive transfer into adult mice. Furthermore, blood cells generated from adult and developmental ECs comparably home to peripheral tissues, where they similarly contribute to inflammatory responses. Thus, our results identify previously unrecognized BM-derived adult hemogenic ECs that generate HSPC and functional mature blood cells.

Sensory neurons drive animal behaviors by detecting environmental stimuli and relaying information to downstream circuits. Beyond their primary roles in sensing, these neurons often form additional synaptic connections outside their main sensory modality, suggesting broader contributions to behavior modulation. Here, we uncover an unexpected role for the thermosensory neuron AFD in coupling tactile experience to locomotion modulation in <i>Caenorhabditis elegans</i>. We show that while AFD employs cyclic guanosine monophosphate (cGMP) signaling for both thermotaxis and tactile-dependent modulation, the specific molecular components of the cGMP pathway differ between these two processes. Interestingly, disrupting the dendritic sensory apparatus of AFD, which is essential for thermotaxis, does not impair tactile-based locomotion modulation, indicating that AFD can mediate tactile-dependent behavior independently of its thermosensory apparatus. In contrast, ablating the AFD neuron eliminates tactile-dependent modulation, pointing to an essential role for AFD itself, rather than its sensory dendritic endings. Further, we find tactile-dependent modulation requires the AIB interneuron, which connects AFD to touch circuits via electrical synapses. Removing innexins expressed in AFD and AIB abolishes this modulation, while re-establishing AFD–AIB connections with engineered electrical synapses restores it. Collectively, these findings uncover a previously unrecognized function of AFD beyond thermosensation, highlighting its influence on context-dependent neuroplasticity and behavioral modulation through broader circuit connectivity.

Acquisition of essential nutrients through diet is crucial for the survival of animals. Dietary odors might enable animals to forage for nutrient-rich diets. We asked if <i>Caenorhabditis elegans</i>, a bacterivorous nematode, uses olfactory cues to forage for essential amino acid-rich (EAA) diets. Using the native microbiota of <i>C. elegans,</i> we show that worms rely on olfaction to select leucine (EAA)-supplemented bacteria. Using gas chromatography, we find that leucine-supplemented bacteria produce isoamyl alcohol (IAA) odor in the highest abundance. Prior adaptation of worms to IAA diminishes the diet preference of worms. Several wild isolates of <i>C. elegans</i> display robust responses to IAA, emphasizing its ecological relevance. We find that foraging for a leucine-supplemented diet is mediated via the AWC olfactory neurons. Finally, we identify SNIF-1 G protein-coupled receptor in AWC neurons as a receptor for IAA and a mediator of dietary decisions in worms. Our study identifies a receptor-ligand module underpinning foraging behavior in <i>C. elegans</i>.

Vestibular hair cells (HCs) convert gravitational and head motion cues into neural signals through mechanotransduction, mediated by the hair bundle—a mechanically integrated organelle composed of stereocilia and a kinocilium. The kinocilium, a specialized form of primary cilium, remains incompletely defined in structure, molecular composition, and function. To elucidate its characteristics, we conducted single-cell RNA sequencing of adult vestibular and cochlear HCs, uncovering a selective enrichment of primary and motile cilia-associated genes in vestibular HCs, particularly those related to the axonemal repeat complex. This enrichment of orthologous axoneme-related genes was conserved in zebrafish and human vestibular HCs, indicating a shared molecular architecture. Immunostaining validated the expression of key motile cilia markers in vestibular kinocilia. Moreover, live imaging of bullfrog and mouse HCs from crista ampullaris revealed spontaneous kinociliary motion. Together, these findings define the kinocilium as a unique organelle with molecular features of primary and motile cilia and suggest its previously unknown role as an active, force-generating element within the hair bundle.

The process of publishing a research article in a scientific journal inevitably involves revising the original version of the article to respond to the concerns raised by peer reviewers. In this article we describe a course module that introduces MSc students at Utrecht University in the Netherlands to this part of the publication process. During the module the students and an invited speaker actively discuss the revision process for a recent article by the speaker. Feedback from students and speakers on the module – which could be readily transferred to other courses in the life and biomedical sciences – has been largely positive.

T cells expressing the γδ T cell receptor (TCR) develop in a stepwise process initiating at the αβ/γδ T cell branch point, followed by maturation and acquisition of effector functions, including the ability to produce interleukin-17 (IL-17) as γδT17 cells. Previous studies linked TCR signal strength and fate choices to the transcriptional regulator HEB (<i>Tcf12</i>) and its antagonist, Id3, but how these factors regulate different stages of γδ T cell development has not been determined. We found that immature fetal γδTCR⁺ cells from conditional <i>Tcf12</i> knockout (HEB cKO) mice were defective in activating the γδT17 program at an early stage, whereas <i>Id3</i>-deficient (Id3-KO) mice displayed a partial block in γδT17 maturation and a defect in IL-17 production. We also found that HEB cKO mice failed to upregulate <i>Id3</i> during γδT17 development, whereas HEB overexpression elevated the levels of <i>Id3</i> in collaboration with TCR signaling. Moreover, Egr2 and HEB were bound to several of the same regulatory sites on the <i>Id3</i> gene locus in the context of early T cell development. Therefore, our findings reveal an interlinked sequence of events during which HEB and TCR signaling synergize to upregulate <i>Id3</i>, which enables maturation and acquisition of the γδT17 effector program.

Motor skill learning is a complex and gradual process that involves the cortex and basal ganglia, both crucial for the acquisition and long-term retention of skills. The cerebellum, which rapidly learns to adjust the movement, connects to the motor cortex and the striatum primarily via the ventral and intralaminar thalamus, respectively. Here, we evaluated the contribution of cerebellar neurons projecting to these thalamic nuclei in a skilled locomotion task in mice. Using a targeted chemogenetic inhibition that preserves the motor abilities, we found that cerebellar nuclei neurons projecting to the intralaminar thalamus contribute to learning and expression, while cerebellar nuclei neurons projecting to the ventral thalamus contribute to offline consolidation. Asymptotic performance, however, required each type of neurons. Thus, our results show that cerebellar neurons belonging to two parallel cerebello-thalamic pathways play distinct, but complementary, roles functioning on different timescales and both necessary for motor skill learning.

Adipose tissues exhibit a remarkable capacity to expand, regress, and remodel in response to energy status. The cellular mechanisms underlying adipose remodelling are central to metabolic health. Hypertrophic remodelling – characterised by the enlargement of existing adipocytes – is associated with insulin resistance, type 2 diabetes, and cardiovascular disease. In contrast, hyperplastic remodelling – in which new adipocytes are generated – is linked to improved metabolic outcomes. Despite its clinical importance, the regulation of hypertrophic and hyperplastic adipose morphology remains poorly understood. Here, we integrate human transcriptomic data with a quantitative CRISPR-imaging platform in zebrafish to identify regulators of adipose morphology. We developed an image-based phenotyping pipeline that captures lipid droplet size, number, and spatial patterning, and applied generalised additive modelling to quantify hyperplastic versus hypertrophic morphology signatures. Using this platform, we conducted an F0 CRISPR screen targeting 25 candidate genes and identified three that induced hypertrophic morphology (<i>txnipa</i>, <i>mmp14b,</i> and <i>foxp1b</i>) and an additional candidate that altered total adiposity (<i>kazna</i>). For functional validation, we generated stable loss-of-function alleles for both zebrafish foxp1 paralogues. Spatial analysis along the anterior-posterior axis revealed that <i>foxp1b</i> mutants display developmental hypertrophy but profoundly blunted adaptive responses to high-fat diet (~68% reduction across all spatial zones), while <i>foxp1a</i> mutants show normal baseline morphology but disrupted spatial patterning of diet-induced hypertrophy. Together, these findings establish a scalable CRISPR-imaging platform for in vivo genetic screening of adipose morphology and reveal distinct roles for Foxp1 paralogues in developmental patterning and adaptive responses to dietary challenge in adipose tissue.

SMC (structural maintenance of chromosomes) protein complexes are ring-shaped molecular motors essential for genome folding. Despite recent progress, the detailed molecular mechanism of DNA translocation in concert with the ATP-driven conformational changes of the complex remains to be clarified. In this study, we elucidated the mechanisms of SMC action on DNA using all-atom and coarse-grained molecular dynamics simulations. We first created a near-atomic full-length model of a prokaryotic SMC–kleisin complex based on experimental structures and implemented ATP-dependent conformational changes using a structure-based coarse-grained model. We further incorporated key protein–DNA hydrogen-bond interactions derived from fully atomistic simulations. Extensive simulations of the SMC complex with 800 base pairs of duplex DNA over the ATP cycle observed unidirectional DNA translocation by the SMC complex. The process exhibited a step size of ~200 base pairs, wherein the SMC complex captured a DNA segment of about the same size within the SMC ring in the engaged state, followed by its pumping into the kleisin ring as ATP was hydrolyzed. Analysis of trajectories identified the asymmetric path of the kleisin as a critical factor for the observed unidirectionality.

The ant nest serves as the skeleton of the ant superorganism. Similar to a skeleton, the nest expands as the colony grows and requires repair after catastrophic events. We experimentally compared nest excavation in colonies seeded from a single mated queen and allowed to grow for 6 months to excavation triggered by a catastrophic event in colonies with fixed demographics, where the age of each worker, including the queen, is known. The areas excavated by equal group sizes differed significantly between these conditions: heterogeneous populations in naturally growing colonies as well as cohorts of young ants dig larger areas than old ant cohorts. Moreover, we find that younger ants tend to dig slanted tunnels while older ants dig straight down. This is a novel form of age polyethism, where an ant’s age dictates not only her likelihood to engage in a task but also the way she performs the task. We further present a quantitative model that predicts that under normal growth, digging is predominantly performed by the younger ants, while after a catastrophe, all ants dig to restore lost nest volume. The fact that the nests of naturally growing colonies exhibit slanted tunnels strengthens this prediction. Finally, our results indicate how a colony’s demographic and physical history are sketched into the current structure of its nest.

Endophilin A3-mediated clathrin-independent endocytosis (EndoA3-mediated CIE) contributes to the internalization of immunoglobulin-like proteins, including key immune synapse components. Here, we identify ICAM1 as a novel EndoA3-dependent cargo, alongside ALCAM. We demonstrate that both proteins subsequently follow retromer-dependent retrograde transport to the <i>trans</i>-Golgi network (TGN) in cancer cells. From there, we propose that they undergo polarized redistribution to the plasma membrane, where they contribute to immune synapse formation between cancer cells and cytotoxic CD8 T cells. Disruption of EndoA3 or retromer components significantly affects the response of autologous cytotoxic CD8 T cells, as evidenced by reduced cytokine production and secretion, but increased lytic activity, while proliferation and later activation marker expression remain intact. This is accompanied by diminished ICAM1 density at the immune synapse, where we observe it arriving via polarized vesicular transport, indicating altered synapse organization. Indeed, cancer cells lacking EndoA3-mediated CIE or retromer form enlarged immune synapses that fail to sustain full T cell cytokine secretion, suggesting a compensatory attempt by T cells to overcome the defective synapse, while likely promoting more transient contacts that potentially favor serial killing. Together, these findings reveal that EndoA3-mediated CIE and retrograde transport act in concert in cancer cells to relocate immune synapse components via the Golgi, thereby fine-tuning the balance between cytotoxic T cell cytokine secretion and lytic activity. These insights contribute to a better understanding of the mechanisms governing immune synapse formation and organization, providing a necessary foundation for the long-term identification of new strategies to enhance T cell–mediated anti-tumor immunity.

Classical psychedelics induce complex visual hallucinations in humans, generating percepts that are coherent at a low level, but which have surreal, dream-like qualities at a high level. While there are many hypotheses as to how classical psychedelics could induce these effects, there are no concrete mechanistic models that capture the variety of observed effects in humans, while remaining consistent with the known pharmacological effects of classical psychedelics on neural circuits. In this work, we propose the ‘oneirogen hypothesis,’ which posits that the perceptual effects of classical psychedelics are a result of their pharmacological actions inducing neural activity states that truly are more similar to dream-like states. We simulate classical psychedelics’ effects via manipulating neural network models trained on perceptual tasks with the Wake-Sleep algorithm. This established machine learning algorithm leverages two activity phases: a perceptual phase (wake) where sensory inputs are encoded, and a generative phase (dream) where the network internally generates activity consistent with stimulus-evoked responses. We simulate the action of psychedelics by partially shifting the model to the ‘Sleep’ state, which entails a greater influence of top-down connections, in line with the impact of psychedelics on apical dendrites. The effects resulting from this manipulation capture a number of experimentally observed phenomena, including the emergence of hallucinations, increases in stimulus-conditioned variability, and large increases in synaptic plasticity. We further provide a number of testable predictions which could be used to validate or invalidate our oneirogen hypothesis.

Genomes assume a complex 3D architecture in the interphase cell nucleus. Yet the molecular mechanisms that determine global genome architecture are only poorly understood. To identify mechanisms of higher-order genome organization, we performed high-throughput imaging-based CRISPR knockout screens targeting 1064 genes encoding nuclear proteins in multiple human cell lines. We assessed changes in the distribution of centromeres at single-cell resolution as surrogate markers for global genome organization. The screens revealed multiple major regulators of spatial distribution of centromeres, including components of the nucleolus, kinetochore, cohesins, condensins, and the nuclear pore complex. Alterations in centromere distribution required progression through the cell cycle and acute depletion of mitotic factors with distinct functions altered centromere distribution in the subsequent interphase. These results identify molecular determinants of spatial centromere organization, and they show that orderly progression through mitosis shapes interphase genome architecture.

Establishing a cell type-specific chromatin landscape is crucial for the maintenance of cell identity during embryonic development. However, our knowledge of how this landscape is set during vertebrate embryogenesis has been limited, due to the lack of methods to jointly detect chromatin modifications and gene expression in the same cell. Here we present a multimodal measurement of full-length transcriptome and histone modifications in individual cells during early embryonic development in zebrafish. We show that before the formation of germ layers, the chromatin and transcription states of cells are uncoupled and become progressively connected during gastrulation and somitogenesis. Silencing of developmental genes is achieved by local spreading of repressive chromatin together with cell type-specific demethylation. Combining transcription factor (TF) expression and chromatin states within an interpretable machine learning model, we classify TFs as lineage-specific activators and repressors and identify a subset of TFs that are epigenetically regulated. Altogether, our data resolves the dynamic relationship between chromatin and transcription during early vertebrate development and clarifies how these two layers interact to establish cell identity.

Protein-protein interactions underlie nearly all cellular processes. With the advent of protein structure prediction methods such as AlphaFold2 (AF2), models of specific protein pairs can be built extremely accurately in most cases. However, determining the relevance of a given protein pair remains an open question. It is presently unclear how to use best structure-based tools to infer whether a pair of candidate proteins indeed interacts with one another: ideally, one might even use such information to screen among candidate pairings to build up protein interaction networks. Whereas methods for evaluating quality of modeled protein complexes have been co-opted for determining which pairings interact (e.g. pDockQ and iPTM), there have been no rigorously benchmarked methods for this task. Here, we introduce PPIscreenML, a classification model trained to distinguish AF2 models of interacting protein pairs from AF2 models of compelling decoy pairings. We find that PPIscreenML outperforms methods such as pDockQ and iPTM for this task, and further that PPIscreenML exhibits impressive performance when identifying which ligand/receptor pairings engage one another across the structurally conserved tumor necrosis factor superfamily (TNFSF). Analysis of benchmark results using complexes not seen in PPIscreenML development strongly suggests that the model generalizes beyond training data, making it broadly applicable for identifying new protein complexes based on structural models built with AF2.

Layer-dependent functional magnetic resonance imaging (fMRI) is a promising yet challenging approach for investigating layer-specific functional connectivity (FC). Achieving a brain-wide mapping of layer-specific FC requires several technical advancements, including sub-millimeter spatial resolution, sufficient temporal resolution, functional sensitivity, global brain coverage, and high spatial specificity. Although gradient echo (GE)-based echo planar imaging (EPI) is commonly used for rapid fMRI acquisition, it faces significant challenges due to the draining-vein contamination. In this study, we addressed these limitations by integrating velocity-nulling (VN) gradients into a GE-BOLD fMRI sequence to suppress vascular signals from the vessels with fast-flowing velocity. The extravascular contamination from pial veins was mitigated using a GE-EPI sequence at 3T rather than 7T, combined with phase regression methods. Additionally, we incorporated advanced techniques, including simultaneous multi-slice (SMS) acceleration and NOise Reduction with DIstribution Corrected principal component analysis (NORDIC PCA) denoising, to improve temporal resolution, spatial coverage, and signal sensitivity. This resulted in a VN fMRI sequence with 0.9 mm isotropic spatial resolution, a repetition time (TR) of 4 s, and brain-wide coverage. The VN gradient strength was determined based on results from a button-pressing task. Using resting-state data, we validated layer-specific FC through seed-based analyses, identifying distinct connectivity patterns in the superficial and deep layers of the primary motor cortex (M1), with significant inter-layer differences. Further analyses with a seed in the primary sensory cortex (S1) demonstrated the reliability of the method. Brain-wide layer-dependent FC analyses yielded results consistent with prior literature, reinforcing the efficacy of VN fMRI in resolving layer-specific functional connectivity. Given the widespread availability of 3T…

The orbitofrontal cortex (OFC) is critical to identifying task structure and to generalizing appropriately across task states with similar underlying or hidden causes. This capability is at the heart of OFCs proposed role in a network responsible for cognitive mapping, and its loss can explain many deficits associated with OFC damage or inactivation. Substance use disorder is defined by behaviors that share much in common with these deficits, such as an inability to modify learned behaviors in the face of new information about undesired consequences. One explanation for this similarity would be if addictive drugs impacted the ability of OFC to recognize underlying similarities, hidden states, that allow information learned in one setting to be used in another. To explore this possibility, we trained rats to self-administer cocaine and then recorded single-unit activity in lateral OFC as these rats performed in an odor sequence task consisting of unique and shared positions. In well-trained controls, we observed chance decoding of sequence at shared positions and near chance decoding even at unique positions, reflecting the irrelevance of distinguishing these positions in the task. By contrast, in cocaine-experienced rats, decoding remained significantly elevated, particularly at the positions that had superficial sensory differences that were collapsed in controls across learning. These neural differences were accompanied by increases in behavioral variability at these positions. A tensor component analysis showed that this effect of reduced generalization after cocaine use also extended across positions in the sequences. These results show that prior cocaine use disrupts the normal identification of hidden states by OFC.

Critical Gram-negative pathogens, like <i>Pseudomonas</i>, <i>Stenotrophomonas,</i> and <i>Burkholderia</i>, are now resistant to most antibiotics. Complex resistance profiles, together with synergistic interactions between these organisms, increase the likelihood of treatment failure in distinct infection settings, for example in the lungs of cystic fibrosis (CF) patients. Here, we discover that cell envelope protein homeostasis pathways underpin both antibiotic resistance and cross-protection in CF-associated bacteria. We find that inhibition of oxidative protein folding inactivates multiple species-specific resistance proteins. Using this strategy, we sensitize multidrug-resistant <i>Pseudomonas aeruginosa</i> to β-lactam antibiotics and demonstrate promise of new treatment avenues for the recalcitrant emerging pathogen <i>Stenotrophomonas maltophilia</i>. The same approach also inhibits cross-protection between resistant <i>S. maltophilia</i> and susceptible <i>P. aeruginosa</i>, allowing eradication of both commonly co-occurring CF-associated organisms. Our results provide the basis for the development of next-generation strategies that target antibiotic resistance, while also impairing specific interbacterial interactions that enhance the severity of polymicrobial infections.

The speed of voluntary movements varies systematically, with some individuals moving consistently faster than others across different actions. These variations, conceptualized as vigor, reflect a time–effort–accuracy tradeoff in motor planning. How do two mechanically coupled partners with different individual vigors collaborate, e.g. to move a table together? Here, we show that such dyads coordinate goal-directed movements with minimal interaction force, exhibiting a <i>dyadic vigor</i> with similar characteristics as individual vigor. The emerging dyadic motor plan is strongly influenced by the slower partner, whose vigor predicts dyadic vigor, with effects lasting beyond practice. Computational modeling with stochastic optimal control reveals the critical role of partners’ movement timing uncertainty and vigor in shaping coordination, allowing us to predict dyadic movements from individual behavior across diverse conditions. These findings shed light on the mechanisms underlying human collaboration and may be used in applications ranging from physical training and rehabilitation to collaborative robotics for manufacturing.

Alpha-band neural oscillations (8–13 Hz) are theorized to phasically inhibit visual processing based, in part, on results showing that pre-stimulus alpha phase predicts detection (i.e., hit rates). However, recent failures to replicate and a lack of a mechanistic understanding regarding how alpha impacts detection have called this theory into question. We recorded EEG while six observers (6020 trials each) detected near-threshold Gabor targets embedded in noise. Using signal detection theory (SDT) and reverse correlation, we observed an effect of occipital and frontal pre-stimulus alpha phase on sensitivity (d'), not criterion. Hit and false alarm rates were counterphased, consistent with a reduction in internal noise during optimal alpha phases. Perceptual reports were also more consistent when two identical stimuli were presented during the optimal phase, suggesting a decrease in internal noise rather than signal amplification. Classification images revealed sharper spatial frequency and orientation tuning during the optimal alpha phase, implying that alpha phase shapes sensitivity by modulating sensory tuning towards relevant stimulus features.

A new biologically-informed training paradigm enables protein language models to predict affinity maturation trajectories for antibodies.

Neural network architecture determines its functional output. However, the detailed mechanisms are not well characterized. In this study, we focused on the neural network architectures of male and hermaphrodite <i>Caenorhabditis elegans</i> and the association with sexually dimorphic behaviors. We applied graph theory and computational neuroscience methods to systematically discern the features of these two neural networks. Our findings revealed that a small percentage of sexual-specific neurons exerted dominance throughout the entire male neural network, suggesting males prioritized sexual-related behavior outputs. Based on the structural and dynamical characteristics of two complete neural networks, sub-networks containing sex-specific neurons and their immediate neighbors, or sub-networks exclusively comprising sex-shared neurons, we predicted dimorphic behavioral outcomes for males and hermaphrodites. To verify the prediction, we performed behavioral and calcium imaging experiments and dissected a circuit that is specific for the increased spontaneous local search in males for mate-searching. Our research sheds light on the neural circuits that underlie sexually dimorphic behaviors in <i>C. elegans</i> and provides significant insights into the interconnected relationship between network architecture and functional outcomes at the whole-brain level.