From work emerging through the middle of the 20th century, the essence of meaning has become widely accepted as being described by the three orthogonal dimensions of valence, arousal, and dominance. These essential dimensions have become the cornerstone of sentiment analysis across many fields. By reexamining first types and then tokens for the English language, and through the use of automatically annotated histograms-"ousiograms"-we find here that the essence of meaning conveyed by words is instead best described by a goodness-power-aggression-danger-structure (GPADS) circumplex framework; that large-scale English language corpora reveal a systematic bias toward safe, low-danger words; and that the power-danger-structure framework is the minimal framework that represents essential meaning. We find remarkable congruences between the GPADS framework and other spaces including mental states and fictional archetypes, and we construct and demonstrate a prototype ousiometer.

Fluorescence microscopy is constrained by optical limits, fluorophore chemistry and finite photon budgets, imposing trade-offs between imaging speed, resolution and phototoxicity. Here we introduce MicroSplit, a deep learning-based computational multiplexing method that enables multiple cellular structures to be imaged simultaneously in a single fluorescent channel and then computationally unmixed. We show that MicroSplit separates up to four superimposed noisy structures into distinct, denoised image channels, enabling faster and more photon-efficient imaging. Built on Variational Splitting Encoder-Decoder networks,  MicroSplit models a posterior distribution over solutions, allowing uncertainty-aware predictions and the estimation of spatially resolved prediction errors from posterior variability. We demonstrate robust performance across diverse datasets, noise levels and imaging conditions, and show that  MicroSplit improves downstream analysis while reducing photon exposure. All methods, data and trained models are released as open resources, enabling immediate adoption of computational multiplexing in biological imaging.

Understanding how nervous systems generate coordinated movement requires precise measurement of body kinematics during natural behavior. The fruit fly, Drosophila, is a model organism with sophisticated behavior and well-studied neural circuits, but tracking fly movements in 3D remains challenging because of their teeny bodies, rapid movements, and frequent self-occlusions. Here we present a pipeline for markerless, full-body 3D pose estimation of fly terrestrial behavior, combining seven synchronized high-speed cameras to capture whole-body kinematics at 800 frames per second. We trained a hybrid 2D/3D deep learning model to track 50 keypoints, then refined them to produce anatomically feasible kinematic trajectories through a retargeting process that solved an inverse kinematics problem constrained by a biomechanical body model. Analysis of 3D kinematics revealed that flies perform grounded running across their full speed range, without transitioning between discrete gaits. Using multi-animal tracking, we found that courting males coordinate both wings during song and modulate body pitch to track the female’s vertical position. Our open-source pipeline and 3D kinematic dataset of fly behavior provide a foundation for neuromechanical modeling and mechanistic studies of motor control in a genetically tractable model organism.

Cells depend on the spatial organization of proteins, RNA, and DNA into discrete subcellular compartments. Previous methods have largely centered on measuring spatial organization based on only one of these biomolecular classes at a time. Here, we demonstrate that POCA photocatalytic proximity labeling can serve as a unified photosensitizer-based platform for profiling the proximal proteomes of protein, RNA, and DNA targets within a single experimental framework. We show that POCA can harness standard immunofluorescence or in situ hybridization workflows to specifically target organic fluorophore photosensitizers to intracellular targets for proximity labeling in fixed cells. POCA-targeted proximity labeling requires minimal cellular input and does not require genetic engineering. Additionally, POCA photosensitizers are selected to also be fluorescent, enabling direct confirmation of on-target localization by imaging prior to proteomic analysis. To demonstrate broad utility, we apply POCA across multiple molecular targets spanning protein, RNA, and genomic DNA, including components of the nuclear pore complex, nucleolus, nuclear speckles, telomeres, and pericentromeric heterochromatin. By anchoring proximity labeling to both a protein and an RNA within the same nuclear compartment, we resolve shared and distinct proximal proteomes from orthogonal molecular perspectives.Competing Interest StatementD.K.S. is a collaborator with Thermo Fisher Scientific, Genentech, Calico Labs, Matchpoint Therapeutics, and AI Proteins. K.M.B is a collaborator with Thermo Fisher Scientific and on the advisory board for Matchpoint Therapeutics. B.J.B. has filed a patent application covering aspects of this work (US Patent App. 18/728,937). B.J.B. is listed as an inventor on patent applications related to the SABER technology related to this work (US Patent 11,492,661; US Patent App. 18/607,269). E.L.H. also collaborates with Thermo Fisher Scientific, Genentech, and Xaira Therapeutics and consults for Calico Labs, Matchpoint Therapeutics, and Flagship Pioneering. Patents and patent applications covering azetidine-containing rhodamine dyes (with inventors J.B.G. and L.D.L.) are assigned to HHMI. L.D.L. is a scientific cofounder, consultant, and shareholder of Eikon Therapeutics. The other authors declare no conflicts.National Institutes of Health, R35GM137916, R35GM150919, DP2GM146246, P30 CA015704, U24HG006673, T32HL007093W. M. Keck Foundation, https://ror.org/000dswa46Pew Charitable Trusts, https://ror.org/02xhk2825Andy Hill CARE FoundationDavid and Lucile Packard Foundation, https://ror.org/032atxq54Damon Runyon Cancer Research Foundation, https://ror.org/01gd7b947

Fructose-1,6-bisphosphate (FBP) is the product of the first committed step of glycolysis, and its concentration is tightly correlated with glycolytic flux. Glycolytic activity varies across tissues and cell types: some tissues, such as the brain, dynamically regulate glycolysis in response to demand, while others, such as the liver have characterized spatial heterogeneity. Here, we report HYlight2, an improved sensor for FBP developed through random whole-gene mutagenesis in E. coli lysate. After four rounds of screening, we isolated HYlight2, which retains its binding affinity while displaying a ΔR/R \~9 in vitro, a three-fold improvement in mammalian cells, and a two-fold improvement in detecting glycolytic responses during stimulated neuronal activity. We further demonstrate its use in vivo to detect altered glycolytic activity in C. elegans neurons, zebrafish pancreatic islets, and mouse liver.

Phagocytosis requires coordinated remodeling of the actin cytoskeleton to generate protrusive and contractile forces that drive target engulfment. Class I myosins Myo1e and Myo1f (Myo1e/f) have been implicated in linking the plasma membrane to the actin network, but their specific roles during Fc-receptor-mediated phagocytosis remain unclear. Using CRISPR-edited RAW 264.7 macrophages lacking Myo1e and Myo1f, we show that double knockout (dKO) cells exhibit markedly reduced uptake of IgG-coated beads, a phenotype that is partially rescued by re-expression of either myosin. Lattice-light-sheet and confocal imaging revealed distinct F-actin architectures corresponding to the various stages of cup progression, including basal podosome-like adhesions, individual phagocytic podosomes (actin teeth) along the rim of the cup, and a contractile phagocytic ring formed by the reorganization of podosomes into a higher-order network. In Myo1e/f- deficient cells, podosome formation was diminished, actin teeth were largely absent, and the phagocytic ring formed prematurely, which was often accompanied by stalled cup progression and repeated engulfment attempts. Myo1e/f localized both to podosomes and to the inner surface of the phagocytic ring, non-muscle myosin II (NM2) localized to the outer surface, and the absence of Myo1e/f correlated with the diffuse distribution of NM2. In addition, Myo1e/f-deficient macrophages exhibited increased trogocytosis of antibody-opsonized HL-60 cells, indicating a shift from whole-target engulfment toward partial target ingestion. These results suggest that Myo1e/f coordinate spatial and temporal transitions between protrusive and contractile actin networks, thereby ensuring efficient phagocytic cup progression. Our findings highlight a dual role for Myo1e/f in adhesion regulation and force balance during macrophage phagocytosis.

In Candida albicans, potassium (K) channels fine-tune ionic balance under stress, contributing to host colonization. Fungal two-pore domain, outwardly rectifying potassium (TOK) channels remain insufficiently characterized despite evidence implicating them in growth and viability. Here, we describe the atomic-resolution structure of a fungal potassium channel, TOK1 from C. albicans (CaTOK), revealing an architecture defined by eight transmembrane helices and a membrane topology distinct from previously characterized K⁺ channel classes. The first four helices form a tetraspanin-like bundle resembling auxiliary subunits of human neuronal ion channels. The pore features an inner helical gating movement analogous to mammalian dimeric K channels, while the K selectivity filter exhibits atypical ion coordination. A cytosolic C-terminal bundle forms an intramolecular network that likely stabilizes CaTOK and may mediate gating. These findings provide a framework for understanding TOK channel function and facilitate future studies of fungal ion homeostasis, pathogenicity, and therapeutic development.

Sensory-guided decisions are the result of sensorimotor transformations across many brain areas. Recent studies have localized the motor- and decision-related components of these transformations using brain-wide neural recordings. It has been more difficult to localize sensory computations in the same way. Here we developed a new approach for linking sensory computations to behavior by training mice to discriminate between two stimuli and testing their responses with new stimuli. In separate animals, we calculated the similarity of neural representations between train and test stimuli, using recordings of up to 73,000 simultaneously-recorded neurons from 9 primary and higher-order visual areas (HVAs) across layers 2 and 3. We found that neural discrimination on test but not train images correlated with behavioral discrimination, and this relation required prior visual experience as it was not present in dark-reared mice. The link between neural and behavioral performance was highest in the medial HVAs, suggesting this region is a critical component of sensory transformations and generalization.

Neural computations are implemented by distributed neural populations that often span multiple brain areas. Causal photo-activation experiments done simultaneously with neural recordings can greatly improve our understanding of these computations, but such methods are typically limited to small subsets of neurons in restricted fields of view. Here we describe a new system called raster photostimulation for photo-activating and recording thousands of neurons, over a short 300 ms time window and over a large 5 mm field-of-view on a two-photon mesoscope. The photo-activation is precisely matched to the neural recording configuration, as it uses the same optical path, although with a different laser that is independently gated. We demonstrate pixel-level precision, frame-by-frame mask updating, and single-frame photostimulation of thousands of neurons. While this method lacks the precise temporal control of alternative methods, it compensates with ease-of-use, spatial precision, cost of implementation and by pushing the limits on the number of near-simultaneously stimulated neurons.

Compositional generation underlies the systematic and essentially unlimited construction of complex concepts from simpler parts, as is foundational to intelligent behavior, but its underlying neural mechanisms remain unclear. Here we reveal a neural implementation of hierarchical compositional construction of abstract sequences. We demonstrate that in an open-ended setting with very sparse feedback, rats innately utilize hierarchical composition to construct adaptive action sequences that would have been difficult to discover from scratch. Prefrontal neural population representations of these abstract sequences adhere to a low-dimensional format that encodes the orderly progression of elemental units comprising the sequence while converging to a sequence-general endpoint. Higher-level compositions in the hierarchy are systematically related to their lower-level constituent parts, reusing much of the representation, while providing context separation and satisfying format constraints. These neural representations are geometrically identical across animals, pointing to a convergent solution for how knowledge is hierarchically assembled via a compositional mechanism.