Biology subjects

Tsutsui, K.

Publications and source records attributed to Tsutsui, K..

13 recordsLinked to original sources

bioRxiv2026

Machine learning-based decoding of emotional valence from electrophysiological signals in the monkey brain

Understanding how the brain operates in naturalistic settings requires methods that go beyond conventional repeated-measurement approaches, necessitating the development of single-trial neural activity analysis. Recent advances in machine learning offer new opportunities for analyzing brain electrophysiological signals. Here, we recorded surface electrocorticography (ECoG) and intracranial local field potentials (LFPs) from emotion-related brain regions in a monkey performing a Pavlovian conditioning task, in which sensory cues predicting reward or punishment were presented randomly, followed by the actual unconditioned outcome. We evaluated the performance of two machine learning algorithms, a Convolutional neural network (CNN) model and a Transformer-based model (EEG-Conformer), in classifying raw ECoG/LFP traces. Both models successfully classified valence type during conditioned and unconditioned stimulus presentation. Furthermore, the Transformer achieved significantly superior classification performance compared to the CNN, particularly in multi-state classification including baseline periods. By optimizing the training dataset for the Transformer model, we could detect dynamic fluctuations in emotional valence consistent with task type from continuously evolving ECoG/LFP patterns recorded throughout the task. These results demonstrate the utility of Transformer-based models for decoding emotional valence from neurophysiological signals in non-human primates.

neuroscience↗

bioRxiv2026

Fragile streak goals induce pressure responses and an inverted-U performance pattern

Psychological pressure is thought to relate to performance in an inverted-U pattern, yet evidence is mixed, possibly because manipulations rarely produce high pressure. We induced scalable pressure using a streak goal that resets after a failure in a force-control task. Participants pursued ten consecutive successes (streak goal) or 100 successes irrespective of sequence (total goal). Under the streak goal, heart rate, pupil size, and perceived pressure rose as participants approached their maximum streak; under the total goal, heart rate and pupil size showed little modulation. Performance followed an inverted-U under the streak goal--improving then declining at the maximum streak--whereas the total goal showed no late-stage drop. This dissociation suggests the late-stage decline reflects pressure, not the streak itself. Despite this clear performance decrement, analyses of movement vigor, feedforward/feedback kinematics, and individual differences in pressure responses revealed no consistent systematic signatures at the group level. Fragile streak goals thus provide a multimodal pressure manipulation and a platform for testing mechanisms underlying choking in human motor control.

neuroscience↗

bioRxiv2025

Individual bounded rationality destabilizes cooperative dynamics in human-AI groups

Cooperative dynamics arise from interactions among individuals, including humans and artificial agents, yet it remains unclear how individual-level cognition and motion shape collective dynamics in such human-AI groups. We combined a dynamic cooperative task, in which one attacker was controlled by a human and two by heuristic artificial agents facing a deep reinforcement learning defender, with counterfactual simulations of artificial groups. Across nine task conditions, groups including a human consistently underperformed relative to groups composed solely of artificial agents. Humans exhibited restricted movement options and probabilistic rather than deterministic action choices, which shifted coordination dynamics towards less symmetric, less organized patterns and reduced performance for both individuals and groups. Embedding these boundedly rational movement and decision parameters into artificial agents reproduced human-like behavior and performance declines. These results demonstrate how limitations of a single individual can destabilize cooperative dynamics and provide insights into the design of robust human-AI systems.

animal behavior and cognition↗

bioRxiv2025

The Paradox of Unpredictability: Trying to Be Unpredictable Makes Sequences More Predictable

Human intuition about randomness is systematically biased. When asked to generate random sequences, people systematically generate too many alternations, yet game theory suggests that being unpredictable to an opponent requires true randomness. In our first experiment (50 Japanese adults), we show that when the goal shifts from being random to being unpredictable, participants do the opposite, generating sequences with a strong repetition bias. In a second experiment, we asked a new sample of 50 Japanese adults to predict two model-derived sequences that were empirically calibrated to the group-level statistics from the first Experiment (Competitive-like vs Random-like). We found a paradox: the repetitive sequences intended to be unpredictable were in fact more predictable than the alternating sequences intended to be random. This paradox appears to stem from an asymmetric reinforcement learning mechanism, whereby success selectively strengthens beliefs in repetition but not alternation. Our findings bridge classic cognitive bias and strategic behavior, suggesting how a fundamental learning mechanism can shape paradoxical predictability.

neuroscience↗

bioRxiv2025

Epigenomic and transcriptomic germ-free ageing atlas reveals sterile inflammation as an intrinsic ageing feature

Inflammageing is a hallmark of ageing. Commensal microbiota plays crucial roles in maintaining tissue homeostasis, yet its impact on cellular ageing and inflammageing remains poorly understood. Here we present a comprehensive single-cell epigenomic and transcriptomic atlas of tissues from mice aged under specific pathogen-free (SPF) or germ-free (GF) conditions. Microbiota conferred beneficial effects in young mice but accelerated various ageing features in old, such as age-related AP-1 pathway upregulation, senescence and transcriptomic alterations, likely due to age-associated dysbiosis. Strikingly, inflammatory signatures persisted across cell types in aged GF mouse tissues, establishing sterile inflammation as an intrinsic feature of ageing. Age-associated B cells expanded equally under GF and SPF conditions, raising the possibility that they function as intrinsic, microbiota-independent drivers of inflammageing and potential therapeutic targets. The atlas provides a resource for distinguishing intrinsic ageing features from those modulated by the microbiota, illuminating mechanisms of cellular ageing and potential anti-ageing interventions. HighlightsSterile inflammation and age-associated B cell expansion are prominent intrinsic ageing features The upregulation of age-associated AP-1 pathways and senescence of alveolar macrophages are attenuated in germ-free condition Germ-free condition induces premature ageing-like features in young mouse tissues but delays ageing features in old mice across multiple cell types

genomics↗

bioRxiv2022

Collaborative hunting in artificial agents with deep reinforcement learning

Collaborative hunting, in which predators play different and complementary roles to capture prey, has been traditionally believed as an advanced hunting strategy requiring large brains that involve high level cognition. However, recent findings that collaborative hunting have also been documented in smaller-brained vertebrates have placed this previous belief under strain. Here, we demonstrate that decisions underlying collaborative hunts do not necessarily rely on sophisticated cognitive processes using computational multi-agent simulation based on deep reinforcement learning. We found that apparently elaborate coordination can be achieved through a relatively simple decision process of mapping between observations and actions via distance-dependent internal representations formed by prior experience. Furthermore, we confirmed that this decision rule of predators is robust against unknown prey controlled by humans. Our results of computational ecology emphasize that collaborative hunting can emerge in various intra- and inter-specific interactions in nature, and provide insights into the evolution of sociality.

animal behavior and cognition↗

bioRxiv2022

Adenosine Reduces Sinoatrial Node Cell AP Firing Rate by uncoupling its Membrane and Calcium Clocks

AO_SCPLOWBSTRACTC_SCPLOWThe spontaneous action potential (AP) firing rate of sinoatrial nodal cells (SANC) is regulated by a system of intracellular Ca2+ and membrane ion current clocks driven by Ca2+-calmodulin-activated adenylyl cyclase-protein kinase A (PKA) signaling. The mean AP cycle length (APCL) and APCL variability inform on the effectiveness of clock coupling. Endogenous ATP metabolite adenosine (ado) binds to adenosine receptors that couple to Gi protein-coupled receptors, reducing spontaneous AP firing rate via G{beta}{gamma} signaling that activates an membrane-clock outward current, IKACh. Ado also inhibits adenylyl cyclase activity via Gi signaling, impacting cAMP-mediated PKA-dependent protein phosphorylation and intracellular Ca2+ cycling. We hypothesize that in addition to IKAdo activation, ado signaling impacts Ca2+ via Gi signaling and that both effects reduce AP firing rate by reducing the effectiveness of the Ca2+ and membrane clock coupling. To this end, we measured Ca2+ and membrane potential characteristics in enzymatically isolated single rabbit SANC. 10 {micro}M ado substantially increased both the mean APCL (on average by 43%, n=10) and AP beat-to-beat variability from 5.1{+/-}1.7% to 7.2{+/-}2.0% (n=10) measured via membrane potential and 5.0{+/-}2.2 to 10.6{+/-}5.9 (n=40) measured via Ca2+ (assessed as the coefficient of variability, CV=SD/mean). These effects were mediated by hyperpolarization of the maximum diastolic membrane potential (membrane clock effect) and suppression of diastolic spontaneous, local Ca2+ releases (LCRs) (Ca2+ clock effect): as LCR size distributions shifted from larger to smaller values, the time of LCR occurrence during diastolic depolarization (LCR period) became prolonged, and the ensemble LCR Ca2+ signal became reduced. The tight linear relationship of coupling between LCR period to the APCL in the presence of ado "drifted" upward and leftward, i.e. for a given LCR period, APCL was prolonged, becoming non-linear indicating clock uncoupling. An extreme case of uncoupling occurred at higher ado concentrations (>100 {micro}M): small stochastic LCRs of the Ca2+ clock failed to self-organize and synchronize to the membrane clock, thus creating a failed attempt to generate an AP resulting in arrhythmia and cessation of AP firing. Thus, the effects of ado to activate G{beta}{gamma} and IKACh, Ado and to activate Gi, suppressing adenylyl cyclase activity, both contribute to the ado-induced increase in the mean APCL and APCL variability by reducing the fidelity of clock coupling and AP firing rate.

physiology↗

bioRxiv2020

Local calcium signals in pacemaker cells heart rate and body mass are self-similar from mice to humans

BackgroundTranslation of knowledge of sinoatrial nodal "SAN" automaticity gleaned from animal studies to human dysrhythmias, e.g. "Sick Sinus" Syndrome (SSS) requiring electronic pacemaker insertion has been sub-optimal, largely because heart rate (HR) varies widely across species. ObjectivesTo discover regulatory universal mechanisms of normal automaticity in SAN pacemaker cells that are self-similar across species. MethodSub-cellular Ca2+ releases, whole cell AP-induced Ca2+ transients and APs were recorded in isolated mouse, guinea-pig, rabbit and human SAN cells. Parametric Ca2+ and Vm Kinetic Transitions (PCVKT) during phases of AP cycles from their ignition to recovery were quantified. ResultsAlthough both action potential cycle lengths (APCL) and PCVKT during AP cycles differed across species by ten-fold, trans-species scaling of PCVKT during AP cycles and scaling, of PCVKT to APCL in cells in vitro, EKG RR intervals in vivo, and BM were self-similar (obeyed power laws) across species. Thus, APCL in vitro, HR in vivo, and BM of any species can be predicted by PCVKT during AP cycles in SAN cells measured in any single species in vitro. ConclusionsIn designing optimal HR to match widely different BM and energy requirements from mice to humans, nature did not "reinvent pacemaker cell wheels", but differentially scaled kinetics of gears that regulate the rates at which the "wheels spin". This discovery will facilitate the development of novel pharmalogic therapies and biologic pacemakers featuring a normal, wide-range rate regulation in animal models and the translation of these to humans to target recalcitrant human SSS. Condensed AbstractStudies in animal models are an important facet of cardiac arrhythmia research. Because HR differs by over ten-fold between some animals and humans, translation of knowledge about regulatory mechanisms of SAN normal automaticity gleaned from studies in animal models to target human SSS has been sub-optimal. Our findings demonstrating that trans-species self-similarity of sub-cellular and cellular mechanisms that couple Ca2+ to Vm during AP cycles can predict heart rate in vivo from mice to humans will inform on the design of novel studies in animal models and facilitate translation of this knowledge to target human disease.

physiology↗

bioRxiv2020

Mapping the molecular and structural specialization of the skin basement membrane for inter-tissue interactions

Inter-tissue interaction is fundamental to multicellularity. Although the basement membrane (BM) is located at tissue interfaces, its mode of action in inter-tissue interactions remains poorly understood, mainly because the molecular and structural details of the BM at distinct inter-tissue interfaces remain unclear. By combining quantitative transcriptomics and immunohistochemistry, we systematically identify the cellular origin, molecular identity and tissue distribution of extracellular matrix molecules in mouse hair follicles, and reveal that BM composition and architecture are exquisitely specialized for distinct inter-tissue interactions, including epidermal-fibroblast, epidermal-muscle and epidermal-nerve interactions. The epidermal-fibroblast interface, namely, hair germ-dermal papilla interface, makes asymmetrically organized side-specific heterogeneity in BM, defined by the newly characterized interface, hook and mesh BMs. One component of these BMs, laminin 5, is required for the topological and functional integrity of hair germ-dermal papilla interactions. Our study highlights the significance of BM heterogeneity in distinct inter-tissue interactions.

developmental biology↗

bioRxiv2020

Synchronized cardiac impulses emerge from multi-scale, heterogeneous local calcium signals within and among cells of heart pacemaker tissue

ABSTRACTBackground The current paradigm of Sinoatrial Node (SAN) impulse generation: (i) is that full-scale action potentials (APs) of a common frequency are initiated at one site and are conducted within the SAN along smooth isochrones; and (ii) does not feature fine details of Ca2+ signalling present in isolated SAN cells, in which small subcellular, subthreshold local Ca2+ releases (LCRs) self-organize to generate cell-wide APs.Objectives To study subcellular Ca2+ signals within and among cells comprising the SAN tissue.Methods We combined immunolabeling with a novel technique to detect the occurrence of LCRs and AP-induced Ca2+ transients (APCTs) in individual pixels (chonopix) across the entire mouse SAN images.Results At high magnification, Ca2+ signals appeared markedly heterogeneous in space, amplitude, frequency, and phase among cells comprising an HCN4+/CX43- cell meshwork. The signalling exhibited several distinguishable patterns of LCR/APCT interactions within and among cells. Apparently conducting rhythmic APCTs of the meshwork were transferred to a truly conducting HCN4-/CX43+ network of straited cells via narrow functional interfaces where different cell types intertwine, i.e. the SAN anatomical/functional unit. At low magnification, the earliest APCT of each cycle occurred within a small area of the HCN4 meshwork and subsequent APCT appearance throughout SAN pixels was discontinuous.Conclusions We have discovered a novel, microscopic Ca2+ signalling paradigm of SAN operation that has escaped detection using low-resolution, macroscopic tissue isochrones employed in prior studies: APs emerge from heterogeneous subcellular subthreshold Ca2+ signals, resembling multiscale complex processes of impulse generation within clusters of neurons in neuronal networks.Condensed abstract By combining immunolabeling with a novel optical technique we detected markedly heterogenous Ca2+signals within and among cell clusters of an HCN4+/CX43- meshwork in mouse sinoatrial node. These Ca2+ signals self-organized and transferred, throughout the node, to projections from an HCN4-/CX43+ network connected to a highly organized, rapidly conducting part of the CX43+ network. Thus, APs emerge from heterogeneous, subthreshold Ca2+ signaling not detected in low-resolution macroscopic isochrones. Our discovery requires a fundamental paradigm shift from concentric impulse propagation initiated within a leading site, to a multiscale/complex process, resembling the emergence of organized signals from heterogeneous local signals within neuronal networks.Competing Interest StatementThe authors have declared no competing interest.Abbreviations listAPaction potentialAPCTAP-induced Ca2+ transientChronopixchrono-pixelCX43Connexin 43 (Gap junction alpha-1 protein)FFTFast Fourier TransformHCN4Hyperpolarization-Activated Cyclic Nucleotide-Gated Channel 4ICaLL-type Ca2+ currentICaTT-type Ca2+ currentIVCInferior Vena CavaLCRLocal Ca2+ ReleaseNCXNa+/Ca2+ exchangerPBSPhosphate-Buffered SalinepreBötCpreBötzinger ComplexROIRegion of InterestSANSinoatrial NodeSVCSuperior Vena CavaView Full Text

physiology↗

bioRxiv2018

Topoisomerase IIβ targets DNA crossovers formed between distant homologous sites to modulate chromatin structure and gene expression

BackgroundType II DNA topoisomerases (topo II) flip the spatial positions of two DNA duplexes, called G- and T-segments, by a cleavage-passage-resealing mechanism. In living cells, these DNA segments can be placed far from each other on the same chromosome. However, no direct evidence for this to occur has been described so far due to lack of proper methodology.\n\nResultsThe beta isoform of topo II (topo II{beta}) is essential for transcriptional regulation of genes expressed in the final stage of neuronal differentiation. To elucidate the enzymes role in the process, here we devise a genome-wide mapping technique for topo II{beta} target sites that can measure the genomic distance between G- and T-segments. It became clear that the enzyme operates in two distinctive modes, termed proximal strand passage (PSP) and distal strand passage (DSP). PSP sites are concentrated around transcription start sites, whereas DSP sites are heavily clustered in small number of hotspots. While PSP represent the conventional topo II targets that remove local torsional stresses, DSP sites have not been described previously. Most remarkably, DSP is driven by the pairing between homologous sequences or repeats located in a large distance. A model-building approach suggested that the DSP sites are intertwined or knotted and topo II{beta} is engaged in unknotting reaction that leads to chromatin decondensation and gene regulation.\n\nConclusionsWhen combined with categorized gene expression analysis, the model-based prediction of DSP sites reveals that DSP is one of the key factors for topo II{beta}-dependency of neuronal gene regulation.

molecular biology↗

bioRxiv2018

Hair Follicle Epidermal Stem Cells Define a Niche for Tactile Sensation

The heterogeneity and compartmentalization of stem cells is a common principle in many epithelia, and is known to function in epithelial maintenance, but its other physiological roles remain elusive. Here we show transcriptional and anatomical contributions of compartmentalized epidermal stem cells (EpSCs) in tactile sensory unit formation in the hair follicle (HF). EpSCs in the follicle upper-bulge, where mechanosensory lanceolate complexes (LCs) innervate, express a unique set of extracellular matrix (ECM) and neurogenesis-related genes. These EpSCs deposit an ECM protein EGFL6 into the collar matrix, a novel ECM that tightly ensheathes LCs. EGFL6 is required for the proper patterning, touch responses, and v integrin-enrichment of LCs. By maintaining a quiescent original EpSC niche, the old bulge, EpSCs provide anatomically stable HF-LC interfaces, irrespective of the stage of follicle regeneration cycle. Thus, compartmentalized EpSCs provide a niche linking the HF and the nervous system throughout the hair cycle.

developmental biology↗

bioRxiv2018

Repulsive Guidance Molecule Acts in Axon Branching in Caenorhabditis elegans

Repulsive guidance molecules (RGMs) are evolutionarily conserved proteins implicated in repulsive axon guidance. Here we report the function of the Caenorhabditis elegans ortholog DRAG-1 in axon branching. The axons of hermaphrodite-specific neurons (HSNs) branch at the region abutting the vulval muscles and innervate these muscles to control egg laying. The drag-1 mutants exhibited defects in HSN axon branching in addition to a small body size and egg laying-defective phenotype. DRAG-1 expression in the hypodermal cells was required for the branching of these axons. The C-terminal glycosylphosphatidylinositol anchor of DRAG-1 was important for its function. Genetic analyses suggested that the membrane receptor UNC-40, but neither SMA-1/{beta}H-spectrin nor SMA-5/MAP kinase 7, acts in the same pathway with DRAG-1 in HSN branching. We propose that DRAG-1 expressed in the hypodermis signals via the UNC-40 receptor expressed in HSNs to elicit branching activity of HSN axons.

developmental biology↗