Lysosomes are essential organelles in eukaryotic cells, required for autophagy, endocytosis, pathogen defense, cell signaling, and metabolic homeostasis. A model of lysosomal ion and water fluxes that captures the synchronized, interdependent operation of ion transporters and diffusion enables prediction of organellar responses to external perturbations and supports the design and interpretation of experiments. Particularly with the advent of organelle-targeted rhodopsin-based optogenetics, there is a pressing need to predict cellular outcomes following light-driven, specific ion transport in lysosomes and other organelles. Currently, no models of lysosomal ion balance fully align with existing experimental data or enable simulation of the organelle's response to stress. Here, we present an updated interactive model that recapitulates appropriate stress responses. We incorporated the functional activities of TPC1 and TMEM165, in addition to the previously included vATPase, ClC-7, TRPML1, and passive ion and water fluxes. The model remains robust during lysosomal maturation, membrane permeabilization, swelling, deacidification induced by vATPase inhibition or additional optogenetics-like proton efflux, and accumulation of weakly basic cationic amphiphilic drugs. Our simulations indicate that lysosomal Ca2+ depletion couples with organellar deacidification triggered by either increased proton leakage or vATPase inhibition, with potential involvement of TMEM165 weakening. Beyond predicting stress-response dynamics, the model enables investigation of highly selective perturbations that can be experimentally induced using optogenetics. Elucidating the mechanisms underlying stable, stress-resilient lysosomal function offers insights for developing anti-disease and antiaging interventions. Further model refinement critically depends on experimental characterization of the lysosomal NHE-like protein mediating sodium influx.
Functional somatic disorders (FSD) are characterized by persistent physical symptoms that cannot be fully explained by structural disease. Whether FSD represent distinct clinical syndromes or a single unified spectrum remains debated. While evidence supports shared epidemiological, psychological, and neurobiological features across syndromes, significant clinical heterogeneity challenges this unified view. This study aimed to identify patient subgroups within an FSD population using data-driven clustering approaches applied to validated self-report measures. Adult patients referred to a specialized FSD consultation at Lille University Hospital (France) between January 2022 and June 2025 were included. Clinical data were collected using validated self-report questionnaires. A Principal Component Analysis (PCA) was performed as an exploratory step to assess data structure. Then, two clustering algorithms were applied directly to scale scores: k-means analysis on continuous data, and Latent Class Analysis (LCA) on dichotomized data. A total of 301 patients were included (66.4% female, mean age 46.7 years). The PCA revealed two main groups of variables, with Childhood Trauma Questionnaire subscales forming a dimension independent of the other scales. Both k-means and LCA identified two groups distinguished by overall symptom severity rather than organ-specific symptom profiles, despite the inclusion of multi-system somatic assessments (Patient Health Questionnaire-15). This suggests that while FSDs present with clinically distinct organ-based symptoms, their underlying psychopathological and somatic burden operates primarily along a severity continuum. Childhood trauma scores were elevated across both clusters, independently of symptom burden. These findings support the hypothesis of a unified FSD spectrum rather than discrete syndromes. Childhood adversity appears to represent a transdiagnostic vulnerability factor independent of clinical severity. Future research should incorporate objective biological markers, longitudinal designs, and broader clinical settings to better capture these clusters and identify predictors of clinical trajectory.
This study presents a bibliometric and scientometric analysis of research trends in schizophrenia genetics over nearly 7 decades (1957-2025). The field has evolved from early heritability studies to genome-wide association studies and multi-omics approaches. The objective was to map the historical trajectory, thematic evolution, key contributors, and global collaboration patterns in relation to clinically relevant psychological constructs. An integrated bibliometric and scientometric approach was applied. A total of 5679 publications were included after systematic retrieval and PRISMA-guided screening from Scopus, PubMed, and Web of Science (initial n = 6193; final n = 5679). Data preprocessing included deduplication, keyword normalization, and metadata verification. VOSviewer (version 1.6.20) was used to construct co-authorship, keyword co-occurrence, and source co-citation networks. Thresholds were defined empirically to balance network interpretability and coverage. The findings suggest a gradual increase in research activity beginning in the 1990s, followed by a marked acceleration after 2010. Four thematic clusters were identified: neurobiological mechanisms and endophenotypes; basic genetic foundations; pharmacological treatments; and clinical comorbidities. Temporal patterns indicate a shift between 2012 and 2016 from genetic association studies toward functional genomics and cognitive neuroscience. Prominent contributors include J. van Os, R.E. Gur, and M.T. Tsuang. The United States may play a prominent role within the global collaboration network. However, clinically relevant psychological constructs may be less centrally integrated within the network structure. This study provides a structured bibliometric mapping of schizophrenia genetics research from 1957 to 2025. The findings may suggest increasing movement toward large-scale, consortium-based research models and a possible partial separation between clinical and basic science domains. Importantly, the results highlight a relative underrepresentation of clinically interpretable psychological dimensions. These findings may indicate the need for stronger integration between genetic discoveries and clinical psychological frameworks to enhance translational relevance.
Attention-Deficit/Hyperactivity Disorder (ADHD), with a prevalence of about 5% in adults, is associated with risky behaviors and increased injuries. This study was conducted to investigate the ADHD prevalence and its related factors among individuals injured in traffic crashes. This population-based cross-sectional study included the drivers who presented to the emergency department due to traffic crash injuries during one year. Data were collected through clinical interviews, review of medical records, and the Conners' Adult ADHD Rating Scales (CAARS) questionnaire. Descriptive statistics were used to summarize demographic characteristics and inferential tests such as chi-square and logistic regression were applied to assess associations between ADHD diagnosis and crash-related variables. 450 drivers injured in traffic crashes in the year 2024 were examined for the prevalence of ADHD. Total prevalence of ADHD among studied drivers was 17.35%. ADHD was more prevalent among younger individuals (p = 0.0235) and males (p = 0.0007). Patients with history of previous crashes were significantly more likely to have ADHD (p = 0.0009). No significant association was found between ADHD and educational level (p = 0.9116), daily (p = 0.443) or weekly (p = 0.076) driving hours, type of vehicle (p = 0.522), or location of the crash (p = 0.825). Based on the main finding the prevalence of ADHD among the studied derivers was 17.35%. The most important related factors of ADHD prevalence were younger age, male gender, and positive history of previous traffic crash injuries.
Aggressive behavior is widely conserved across animal species and is often reduced during domestication. Recent advances in neurotechnology using adeno-associated virus (AAV) have enable identification of neural circuits underlying aggression in rodents, whereas those in domesticated birds remain poorly understood. Here, we applied a chemogenetic strategy using AAV-mediated expression of designer receptors exclusively activated by designer drugs (DREADDs) in chickens. Immunostaining of c-Fos following fight showed neuronal activation in the amygdala core nucleus (ACo), the amygdala intermedioanterior nucleus, and the ventromedial hypothalamic nucleus. AAV successfully transduced DREADDs in the chicken ACo. The frequency of aggressive behavior was higher under deschloroclozapine treatment condition than under dimethyl sulfoxide control at the first and second administrations, whereas this effect returned to baseline levels by the final administration. These findings suggest brain region-specific chemogenetic modulation of aggression in chickens and provide new insight into the neural basis of social behavior in domesticated animals.
In the developing human brain, forebrain pericytes of neural crest origin are proposed to exert their function at the leading front of growing vessels. The combined sprouts are made up by an endothelial tip cell and a more advanced pericyte extended tunnelling nanotube (TNT), which forms 'intervascular bridges' during vessel branching. To better understand of the process of pericyte TNTs (P-TNTs)-guided vessel growth, we have immunolocalized the c-KIT (CD117) receptor and its ligand SCF, as a possible 'non canonical' signaling pathway involved in endothelium-pericyte interactions during the pericyte-driven mode of sprouting. Immunofluorescent high-resolution confocal microscopy and 3D modelling were applied on human developing brain and glioblastoma (GB) sections. TNTs were revealed by sequential scanning of 20-µm tissue sections, leveraging on NG2 proteoglycan as a marker of immature/re-activated pericytes and on collagen type IV (COL IV) as a TNT-associated ECM molecule. Confocal-aided morphometry was applied to evaluate the density of NG2+ TNTs on brain and tumor vascular networks. The cellular and subcellular immunolocalization of c-KIT and SCF was achieved by double staining with NG2 and the endothelial cell-marker CD31. In both developing brain and GB samples, P-TNTs are seen associated with pericytes budding from parental vessels or advancing at the distal edge of combined vessel sprouts. The number of TNTs largely prevails in fetal brain, compared with peripheral GB areas, while it significantly reduces in tumor core areas. c-KIT+/SCF+ endothelial cells lie close to c-KIT-/SCF+ pericytes in both fetal and tumor samples. In endothelial cells of fetal brain microvessels, c-KIT shows a dual cell membrane and nuclear localization, the latter being barely detectable on endothelial cells of tumor vessels. P-TNTs, which are tightly associated in situ with endothelial-pericyte combined sprouts, appear to play a dual role during vessel collateralization by bridging the gap between distant vessels and guiding vascular outgrowth. The complementary cellular distribution of c-KIT and SCF observed in endothelial cells and pericytes suggests that both endothelial autocrine/paracrine SCF/c-KIT signaling and pericyte-derived paracrine/juxtacrine SCF cues may contribute to the pericyte-driven mode of vessel branching. Similar observations in GB samples further suggest a potential involvement of pericytes and their P-TNTs in tumor vascularization, although sprouting endothelial cells displayed distinct subcellular patterns of c-KIT expression in fetal versus GB tissues.
Autism spectrum disorder (ASD) is characterized by alterations in social understanding and self-related experience that overlap with broader dimensions of psychosocial vulnerability. These domains are tightly interconnected, motivating the use of analytic approaches that can capture their organization as complex associations rather than as isolated dimensions. We applied network analysis to autistic adults (N = 156) and demographically matched neurotypical controls (N = 454). Gaussian graphical models were estimated using sum scores of psychosocial constructs, and networks were compared using centrality metrics and the network comparison test. Compared to neurotypical participants, autistic individuals showed higher levels of psychosocial difficulties and lower global network strength. Mentalization emerged as a central node in both groups, while autistic traits were more central in the neurotypical network and trait anxiety showed relatively higher centrality in the ASD network. These findings suggest that psychosocial vulnerability in autism is characterized by a distinct and less integrated network organization, with mentalization playing a central role across groups and anxiety showing a relatively greater centrality in ASD. Network-based approaches may therefore help identify mechanism-relevant targets for intervention and refine dimensional models of social and self-related functioning in autism. Autism affects how people understand themselves and others, including social relationships, emotions, and everyday stress. These experiences are closely connected rather than independent. We studied how these different experiences relate to each other in autistic and non‐autistic adults using a method that looks at patterns of connections rather than single traits. In autism, these connections were less tightly linked, suggesting a more fragmented organization of psychosocial experiences. Across both groups, understanding mental states (mentalization) played a key role, while anxiety had a stronger influence in autism, highlighting possible targets for support and intervention.
Pathogenic variants in GBA1 are a significant genetic risk factor for Parkinson disease (PD). Owing to high sequence homology between GBA1 and its pseudogene GBAP1, short-read NGS (srNGS) is susceptible to read misalignment, particularly in the presence of recombinant alleles. A targeted short-read NGS (srNGS) panel was applied to 175 selected Korean patients with PD, including those with early-onset disease, family history, preoperative assessment, or atypically rapid progression, identifying GBA1 copy-number loss in five patients. Confirmatory testing using MLPA, long-range PCR followed by Sanger sequencing, and long-read sequencing demonstrated that these findings represented recombinant alleles rather than simple deletions in four cases, while confirmatory analysis could not be performed in one case due to insufficient DNA. Thus, recombinant alleles were identified in at least 2.3% (4/175) of patients. We further evaluated whether GBAP1 copy-number analysis could aid interpretation of srNGS findings. GBAP1 copy-number patterns were concordant with confirmatory results in all four cases, indicating gene conversion events; in the remaining case, for which confirmatory testing could not be performed due to insufficient DNA, the pattern was suggestive of a fusion allele. Our findings demonstrate that recombinant alleles may present as exon-level deletions on targeted srNGS. We propose that concurrent GBAP1 copy-number analysis provides a practical interpretive aid for distinguishing recombinant rearrangements from simple copy-number alterations and for guiding confirmatory test selection in routine diagnostics.
Elucidating the relationships among in vivo activity, brain-wide projection, and gene expression is critical for understanding neuronal functions, but characterizing these modalities for the same neuron remains technically challenging. Here, we developed a trimodal platform combining in vivo Ca2+ imaging, morphological reconstruction of single neurons in cleared whole brains, and post hoc imaging-based in situ transcriptomic profiling in thick brain sections. We applied this platform to the mouse primary visual cortex (VISp) and obtained trimodal profiles for 141 intratelencephalic (IT) and pyramidal tract (PT) neurons. We found that regional axonal arborization, soma location, transcriptomic signatures, and subcellular RNA localization emerged as informative predictors for distinguishing neurons preferentially responsive to different visual stimuli. Importantly, morphological and transcriptomic features are complementary and, when integrated, can better predict neuronal function. Thus, this trimodal platform enables a comprehensive understanding of the relationships among gene expression, morphological diversity, and functional properties of single neurons.
Rape myth acceptance has been connected to increased victim blaming and reduced perpetrator blame. However, variability exists in how scholars conceptualize relationships between rape myth stereotypes and assignment of victim and perpetrator blame. The current study sought to examine how social dominance orientation and personal and social power perceptions predict blame assignment in sexual assault scenarios among a community sample recruited from social media (N = 462). Social dominance orientation significantly predicted greater victim blame and lower perpetrator blame after controlling for sex assigned at birth, prior victimization, and knowledge of others' victimization. Findings have applied implications for therapeutic interventions and system-level prevention.
The mammalian central circadian clock resides in the suprachiasmatic nucleus (SCN) of the hypothalamus in the brain and is responsible for coordinating daily rhythms of biological processes spanning from gene expression to behavior. Light, the primary environmental zeitgeber, entrains the SCN via melanopsin-expressing intrinsically photosensitive retinal ganglion cells that project through the retino-hypothalamic tract. Altered circadian rhythms are common in individuals diagnosed with neurodevelopmental and neurodegenerative disorders, and often, associated with structural alterations of the SCN and impaired retinal input; importantly, these anomalies can be recapitulated in animal models. Here, we describe step-by-step protocols for quantitative histomorphometrical analysis of the SCN and the assessment of retinal-SCN connectivity, previously used in mouse models of neurodevelopmental and neurodegenerative disorders. These include measurement of the SCN area, perimeter, height and width using Nissl- or DAPI-stained coronal sections, as well as densitometric and plot profile analyses of cholera toxin β-subunit-labeled retinal projections using Axiovision or Fiji/ImageJ. The protocols incorporate standardized region-of-interest, measurements by masked observers, and consistent scaling procedures to enhance reproducibility. These methods provide a rigorous framework for detecting structural anomalies and connectivity defects in the circadian system and can be broadly applied to other experimental models of circadian dysfunction. Key features • Histomorphometrical analyses of the SCN can provide anatomical bases to understand altered sleep and circadian rhythms in animal models of disease. • Exploration of retinal-SCN connectivity to facilitate the identification of the underlying causes of deficits in the response to photic cues in animal models of disease. • The protocols described here employ widely used and accessible software and provide rigorous but easy-to-follow instructions. • These analyses do not require expensive staining procedures and can be easily implemented in any laboratory. • Strengths for reproducibility: usage of fixed region-of-interest (ROI), measurements averaged from multiple sections per animal, masked observers thoroughly trained.
A variety of rating scales are currently being used to assess symptom severity and quantify symptoms change in attention-deficit/hyperactivity disorder (ADHD) research and clinical practice. This poses difficulties in interpreting scores from different scales in clinical practice and synthesizing data from studies using different scales. We aimed to develop algorithms for converting scores across the ADHD scales most often used in randomized controlled trials (RCTs) of ADHD medications in children/adolescents and adults, and to develop an online tool for implementing the algorithms. We analyzed individual participant data from RCTs of ADHD medications (32 RCTs in children/adolescents, 21 in adults), with data on at least two scales per participant at the same timepoint. We applied a series of competing models, that is, univariable and multivariable regression, random forests, and an equipercentile linkage approach, to link pairs of scales. To assess the error of the linking procedure and identify the optimal model, we calculated the median absolute error and R2 of all approaches by comparing the values predicted from the models to the observed ones. We subsequently developed a tool to implement the best algorithms. We linked six commonly used ADHD scales, such as the ADHD Rating Scale (ADHD-RS-IV; investigator-rated) and the Conners' Parent Rating Scale (CPRS-R:S). Spline models most frequently yielded the lowest prediction error, outperforming alternative conversion algorithms for absolute scores in 6 out of 12 univariable models and 8 out of 12 multivariable models. The tool for scores conversion is available at ADHD_Scale_Conversion_Tool. Our linkage algorithms enable the comparison and harmonization of findings across studies using different ADHD rating scales. Translating scores across scales improves the interpretability of research findings, facilitates future evidence synthesis across studies, and may support clinical practice. Our online tool supports the practical uptake of our results.
Magnetoelectric nanoparticles (MENPs) represent a wireless neuromodulation strategy capable of converting externally applied magnetic fields into localized electric stimulation. Although MENPs have shown neurostimulatory effects in cellular and rodent models, their ability to activate viable adult human brain tissue remains insufficiently explored. This study evaluated whether MENPs can induce neuronal activation in ex vivo adult human cortical brain slices. Organotypic human brain slice cultures were prepared from resected temporal lobe tissue obtained during epilepsy surgery. Slices were transduced with a neuron specific GCaMP7 calcium indicator and microinjected with CoFe2O4-BaTiO3 MENPs. Magnetic stimulation was delivered using a 6 mT, 140 Hz alternating magnetic field superimposed on a 220 mT direct-current bias. Calcium imaging was performed before, during, and after stimulation. Control conditions included sham stimulation, reversed coil stimulation, field only and MENPs only conditions, and calcium-channel blockade with CdCl2 or NNC 55-0396. MENPs implanted slices exposed to the magnetic field showed increased neuronal calcium activity, reflected by a higher percentage of significantly responsive regions of interest during and/or after stimulation. Comparable activation was not observed in MENPs-only (sham controls) or field-only controls. Reversed coil stimulation reduced the response. Calcium-channel blockade attenuated stimulation-associated activity, supporting involvement of voltage-gated calcium dependent mechanisms. MENPs can evoke neuronal activity in viable adult human cortical tissue ex vivo. These findings provide translational evidence supporting MENPs as a potential wireless neuromodulation platform and establish human brain slice cultures as a relevant model for preclinical evaluation of nanoparticle mediated neurostimulation.
Randomized trials support early initiation of direct oral anticoagulants (DOACs) after atrial fibrillation (AF)-related ischemic stroke, but patients with breakthrough ischemic stroke occurring despite ongoing anticoagulation have been largely under-represented. We evaluated the effectiveness and safety of early versus delayed DOAC initiation after breakthrough ischemic stroke. We performed a target trial emulation comparing early versus delayed DOAC initiation in patients with breakthrough ischemic stroke. Treatment strategies were prespecified using severity-adapted timing thresholds based on baseline National Institutes of Health Stroke Scale (NIHSS) scores. The study population was drawn from the retrospective arm of the international, multicentre ASPERA study and included patients with AF who experienced an ischemic stroke while receiving continuous anticoagulation. To emulate random assignment and avoid immortal time bias, a cloning-censoring-weighting approach with inverse probability weighting was applied. Primary outcomes were 90-day new ischemic events and moderate-to-severe bleeding. Risk ratios (RRs), absolute risk differences (RDs), and hazard ratios (HRs) were estimated using weighted regression and Cox models. Among 833 patients (median age 81 years), 336 were assigned to early and 497 to delayed DOAC initiation. At 90 days, early initiation was associated with a lower risk of new ischemic events (RR 0.44, 95% CI 0.21-0.90; RD -3.64%, 95% CI -6.40 to -0.87; HR 0.43, 95% CI 0.21-0.91). Moderate-to-severe bleeding occurred less frequently with early initiation (RR 0.10, 95% CI 0.01-0.76). Early initiation was also associated with lower 90-day all-cause and vascular mortality. A Net Early Benefit Score integrating ischemic and bleeding risks was positive across all NIHSS strata. In patients with breakthrough ischemic stroke, early severity-adapted DOAC initiation was associated with lower risks of recurrent ischemic events and mortality at 90 days without an increase in major bleeding. These findings support early anticoagulation initiation in this high-risk population.
Older adults experience a high risk of falls, often due to impaired stepping responses that require rapid and efficient weight transfer. This pilot study examined whether age group and principal component analysis (PCA)-derived profiles of muscle morphology, muscle quality, and strength were associated with weight-transfer performance during voluntary forward and backward stepping. Twenty-three younger and older adults underwent ultrasound imaging of the tensor fasciae latae (TFL), vastus lateralis (VL), and biceps femoris (BF), assessments of knee extensor, knee flexor, and hip abductor strength, and completion of a Choice Reactive Stepping Test (CRST). PCA was applied to nine muscle-related variables to summarize interrelated measures of muscle size, echo intensity, and strength into neuromuscular profiles. The first four principal components (PCs) explained approximately 80.8% of the total variance. PC1 reflected a favorable muscle quality and strength profile, characterized by lower echo intensity and greater hip abduction and knee flexion strength. PC2 suggested a dissociation between TFL thickness and force-generating capacity, whereas PC3 and PC4 reflected muscle-specific structural and strength heterogeneity. PCA-derived PC scores and age group were then entered into a multivariate multiple regression (MMR) model, with weight-transfer onset (WTO) and weight-transfer duration (WTD) during forward and backward stepping as response variables. The MMR model did not provide statistical evidence that age group or any PC was associated with the combined WTO and WTD outcomes. These findings suggest that, in this sample, resting muscle morphology, muscle quality, and maximal strength did not explain weight-transfer performance when modeled as integrated neuromuscular profiles. Weight-transfer performance during stepping may depend on dynamic neuromuscular factors, such as rapid force generation, activation timing, coordination, and task-specific balance strategies, which should be examined in larger studies.
How do we recognize emotions in others' faces? Embodied-simulation theory suggests we mimic observed expressions to enhance our understanding. However, empirical evidence for this online mimicry account remains mixed. Prior research typically measured facial mimicry and emotion recognition concurrently. Here, we test a complementary embodied-simulation account by measuring deliberate facial mimicry quality as an independent skill. We test whether individual differences in this skill - potentially reflecting richer sensorimotor representations - predict emotion recognition performance. In Study 1 (N = 34), participants completed three facial-emotion-recognition tasks, then performed an instructed-mimicry task. We developed and validated a deep-learning-based algorithm to quantify mimicry accuracy and timing, applied diffusion modeling to capture recognition efficiency, and linked mimicry skill to emotion recognition performance. Study 2 (N = 50) was a preregistered replication. Results showed that accurate mimickers recognized emotions more precisely and efficiently. Additionally, longer mimicry latencies predicted higher recognition accuracy. Our findings link facial mimicry quality and emotion recognition performance, highlighting mimicry as a measurable skill that predicts social-cognitive ability. The online version contains supplementary material available at 10.1007/s42761-026-00366-9.
Heart rate variability (HRV) has been proposed as a marker of autonomic dysfunction in somatic symptom disorder (SSD), but findings based on single parameters remain inconsistent. Pattern-based approaches may provide a more integrative characterization of autonomic regulation. This study applied a four-pattern HRV classification (normal, low, relatively high sympathetic, relatively high vagal) in 148 patients with SSD from the SOMA.SSD cohort. Resting-state HRV was assessed using 5-min ECG recordings at baseline. Psychopathology (PHQ-15, PHQ-9, SSD-12) was assessed at baseline, 6, and 12 months. Group differences and longitudinal effects were analyzed using ANOVA and linear mixed-effects models. Overall, no significant differences in symptom levels were observed across the four HRV patterns groups at baseline. However, post-hoc comparisons indicated that patients with a low HRV pattern reported higher somatic symptom severity, depressive symptoms, and SSD-related distress compared to the normal HRV group. These differences remained stable over 12 months, with no significant time or interaction effects. Relatively high sympathetic and relatively high vagal patterns were not associated with increased psychopathology. A pattern-based HRV classification is applicable in patients with SSD and highlights the clinical relevance of a low HRV pattern. This approach may support physiologically informed subtyping in psychosomatic research and could have implications for future clinical stratification.
Many dynamical systems, ranging from genetic circuits to the human brain to human social systems, are often characterized as computational. Although extensive research has explored their dynamics, the computations underlying often remain elusive. Even the fundamental task of quantifying the amount of computation underlying a dynamical system remains underinvestigated. In this study we introduce a task-independent framework to estimate the amount of computation implemented by an observed system based on empirical time-series of its dynamics. This framework works by forming a statistical reconstruction of that dynamics, and defining the amount of computation in terms of both the complexity and fidelity. We validate our framework by showing it appropriately distinguishes the relative amount of computation across different regimes of Lorenz dynamics and various computation classes of cellular automata. We then apply this framework to whole-brain neural recordings of Caenorhabditis elegans and large scale population recordings of the mouse cortex. We find that high and low amounts of computation underlie the neural dynamics of freely moving and immobile worms. Our analysis further sheds light on the amount of computation C. elegans performs in various locomotion states. When applied to large-scale electrophysiological recordings from the mouse cortex during a visual decision-making task, our framework recovers the ground-truth difficulty of the task, assigning higher amounts of computation to more difficult trials where sensory inputs are ambiguous. In sum, our study explores a powerful framework for quantifying the amount of computation performed by a system based on time-series data of its dynamics, and highlights neural computation in both simple and complex organisms.
Stochastic differential equations (SDEs) are effective tools for modelling various real-world phenomena, ranging from chemical reactions to neural dynamics. In this paper, we propose a flow-matching-based SDE learning framework, called FlowSDE, to model the non-equilibrium dynamics and identify its critical state transitions. FlowSDE combines conditional flow matching and physical constraints to construct data-driven representations of the system's evolution, capturing mean-field potentials and detecting transition points. We validate FlowSDE with various dynamical systems, demonstrating its ability to uncover latent transition points and predict transitions with greater accuracy. We also compare FlowSDE with traditional early warning signals (e.g. variance, lag-1 autocorrelation (lag-1 AC)) in epileptor models, highlighting the robustness of our approach under non-stationary or noisy conditions. Overall, the FlowSDE framework offers an interpretable, flexible and robust solution to characterize and forecast complex dynamics, with a wide range of potential applications in neuroscience and other fields. This article is part of the theme issue 'Critical transitions and intelligent control in complex systems'.
High-throughput genomic and proteomic technologies are used to study biological systems by performing differential expression analysis across various experimental conditions. Geneset Ordinal Association Test (GOAT) is an analytic method recently introduced to statistically evaluate the differential expression of a defined set of genes or proteins. Despite the availability of numerous enrichment tools, many lack accessibility for users without programming expertise, provide limited continuation beyond listing top enriched terms, and offer little support for interactive visual exploration or hypothesis generation. Moreover, existing web-based platforms rarely support multi-contrast comparisons and generally omit gene-level or network-based context for pathway analysis. To address these limitations, we present Geneset Ordinal Association Test Enrichment Analysis (GOATEA), an R/Shiny application that implements and extends the GOAT algorithm with interactive visualization, multi-contrast comparison, and integrated gene- and network-based context for bottom-up pathway analysis, enabling comprehensive enrichment analysis. GOATEA supports independent analysis of transcriptomic and proteomic data. To demonstrate its capability to integrate matched modalities, we applied it to the Colameo dataset containing paired mass spectrometry and RNA sequencing data. This proof-of-concept example highlights the tool's strength in enabling multi-omics analyses and simultaneous comparison of multiple contrasts. An interactive overlap analysis identified 458 shared genes for focused enrichment and network exploration. By integrating these results in a gene- and network-based context for bottom-up pathway analysis, GOATEA applies a stringent interaction confidence threshold to emphasize qualitative protein-protein interactions, highlighting topic-relevant associations for further hypothesis generation. GOATEA streamlines enrichment analysis workflows by combining the GOAT algorithm with interactive visualizations in a user-friendly graphical interface. It facilitates exploratory analysis and hypothesis generation for researchers with or without programming expertise. GOATEA is available as an open-source tool, with full documentation, including usage vignettes (https://mauritsunkel.github.io/goatea/).