Patient healthcare utilization consists of irregularly time-stamped events, such as outpatient visits, inpatient admissions, and emergency encounters, forming individualized care trajectories. Modeling these trajectories is crucial for understanding utilization patterns and predicting future care needs, but is challenging due to temporal irregularity and severe class imbalance. In this work, we build on the Transformer Hawkes Process framework to model patient trajectories in continuous time. By combining Transformer-based history encoding with Hawkes process dynamics, the model captures event dependencies and jointly predicts event type and time-to-event. To address extreme imbalance, we introduce an imbalance-aware training strategy using inverse square-root class weighting. This improves sensitivity to rare but clinically important events without altering the data distribution. Experiments on real-world data demonstrate improved performance and provide clinically meaningful insights for identifying high-risk patient populations.

Computer vision methods are increasingly used for the automated analysis of large volumes of video data collected through camera traps, drones, or direct observations of animals in the wild. While recent advances have focused primarily on detecting individual actions, much less work has addressed the detection and annotation of interactions -- a crucial aspect for understanding social and individualized animal behavior. Existing open-source annotation tools support either behavioral labeling without localization of individuals, or localization without the capacity to capture interactions. To bridge this gap, we present SiLVi, an open-source labeling software that integrates both functionalities. SiLVi enables researchers to annotate behaviors and interactions directly within video data, generating structured outputs suitable for training and validating computer vision models. By linking behavioral ecology with computer vision, SiLVi facilitates the development of automated approaches for fine-grained behavioral analyses. Although developed primarily in the context of animal behavior, SiLVi could be useful more broadly to annotate human interactions in other videos that require extracting dynamic scene graphs. The software, along with documentation and download instructions, is available at: https://silvi.eckerlab.org.

Amortized Bayesian inference (ABI) with neural networks has emerged as a powerful simulation-based approach for estimating complex mechanistic models. However, extending ABI to hierarchical models, a cornerstone of modern Bayesian analysis, has been a major hurdle due to the need to simulate and process massive datasets. Our study tackles these challenges by extending compositional score matching (CSM), a divide-and-conquer strategy for Bayesian updating using diffusion models. We develop a new error-damping estimator to address previous stability issues of CSM when aggregating large numbers of data points. We first verified the numerical stability with up to 100,000 data points on a controlled benchmark. We then evaluated our method on a hierarchical AR model, achieving competitive performance to direct ABI baselines on smaller problem sizes while using less than one full model simulation for larger problem sizes. Finally, we address a large-scale inverse problem in advanced microscopy with over 750,000 parameters, demonstrating its relevance to real scientific applications.

Logical modeling is a powerful tool in biology, offering a system-level understanding of the complex interactions that govern biological processes. A gap that hinders the scalability of logical models is the need to specify the update function of every vertex in the network depending on the status of its predecessors. To address this, we introduce in this paper the concept of strong regulation, where a vertex is only updated to active/inactive if all its predecessors agree in their influences; otherwise, it is set to ambiguous. We explore the interplay between active, inactive, and ambiguous influences in a network. We discuss the existence of phenotype attractors in such networks, where the status of some of the variables is fixed to active/inactive, while the others can have an arbitrary status, including ambiguous.

Schizophrenia (SZ) is a brain disorder leading to detached mind's normally integrated processes. Hence, the exploration of the symptoms in relation to functional connectivity (FC) had great relevance in the field. FC can be investigated on different levels, going from global features to single edges between regions, revealing diffuse and localized dysconnection patterns. In this context, SZ is characterized by a diverse global integration with reduced connectivity in specific areas of the Default Mode Network (DMN). However, the assessment of FC presents various sources of uncertainty. This study proposes a multi-level approach for more robust group-comparison. FC between 74 AAL brain areas of 15 healthy controls (HC) and 12 SZ subjects were used. Multi-level analyses and graph topological indexes evaluation were carried out by the previously published SPIDER-NET tool. Robustness was augmented by bootstrapped (BOOT) data and the stability was evaluated by removing one (RST1) or two subjects (RST2). The DMN subgraph was evaluated, toegether with overall local indexes and connection weights to enhance common activations/deactivations. At a global level, expected trends were found. The robustness assessment tests highlighted more stable results for BOOT compared to the direct data testing. Conversely, significant results were found in the analysis at lower levels. The DMN highlighted reduced connectivity and strength as well as increased deactivation in the SZ group. At local level, 13 areas were found to be significantly different ($p<0.05$), highlighting a greater divergence in the frontal lobe. These results were confirmed analyzing the negative edges, suggesting inverted connectivity between prefronto-temporal areas. In conclusion, multi-level analysis supported by BOOT is highly recommended, especially when diffuse and localized dysconnections must be investigated in limited samples.

Bacteriophages, often referred to as the dark matter of the biosphere, play a critical role in regulating microbial ecosystems and in antibiotic alternatives. Thus, accurate interpretation of their genomes holds significant scientific and practical value. While general-purpose Large Language Models (LLMs) excel at understanding biological texts, their ability to directly interpret raw nucleotide sequences and perform biological reasoning remains underexplored. To address this, we introduce PhageBench, the first benchmark designed to evaluate phage genome understanding by mirroring the workflow of bioinformatics experts. The dataset contains 5,600 high-quality samples covering five core tasks across three stages: Screening, Quality Control, and Phenotype Annotation. Our evaluation of eight LLMs reveals that general-purpose reasoning models significantly outperform random baselines in phage contig identification and host prediction, demonstrating promising potential for genomic understanding. However, they exhibit significant limitations in complex reasoning tasks involving long-range dependencies and fine-grained functional localization. These findings highlight the necessity of developing next-generation models with enhanced reasoning capabilities for biological sequences.

Large Language Models (LLMs) are increasingly adopted as conversational assistants in genomics, where they are mainly used to reason over biological knowledge, annotations, and analysis outputs through natural language interfaces. However, existing benchmarks either focus on specialized DNA models trained for sequence prediction or evaluate biological knowledge using text-only questions, leaving the behavior of general-purpose LLMs when directly exposed to raw genome sequences underexplored. We introduce GenomeQA, a benchmark designed to provide a controlled evaluation setting for general-purpose LLMs on sequence-based genome inference tasks. GenomeQA comprises 5,200 samples drawn from multiple biological databases, with sequence lengths ranging from 6 to 1,000 base pairs (bp), spanning six task families: Enhancer and Promoter Identification, Splice Site Identification, Taxonomic Classification, Histone Mark Prediction, Transcription Factor Binding Site Prediction, and TF Motif Prediction. Across six frontier LLMs, we find that models consistently outperform random baselines and can exploit local sequence signals such as GC content and short motifs, while performance degrades on tasks that require more indirect or multi-step inference over sequence patterns. GenomeQA establishes a diagnostic benchmark for studying and improving the use of general-purpose LLMs on raw genomic sequences.

This chapter is an overview of foundational results in the mathematical theory of replicator systems. Its primary aim is to provide a unified framework for the mathematical formalisation of evolutionary processes in the spirit of generalised Darwinism -- that is, for any system in which heredity, variability, and selection can be meaningfully defined, regardless of the specific biological substrate. Starting from the Kolmogorov equations for interacting populations, we derive the replicator equation and examine three canonical regimes: independent, autocatalytic, and hypercyclic replication. The hypercycle is shown to be permanent and to carry evolutionary variability intrinsically. We then survey the quasispecies framework -- the Eigen and Crow--Kimura models -- covering global stability of equilibria, sequence space structure, and the error-threshold phenomenon. Throughout, the emphasis is on the mathematical structures that underlie these models rather than on biological detail, with the goal of making the framework applicable to abstract evolutionary dynamics beyond its original molecular biology context.

The geometry of blood vessels strongly affects hemostasis and thrombosis through red blood cell (RBC) dynamics and platelet margination. Growing platelet aggregates, in turn, reshape the local vessel wall topography, leading to a strongly coupled system. However, it is not well understood how surface heterogeneities alter local hemodynamics and platelet margination, thereby driving further aggregate growth. This study investigates how hematocrit (Ht) and shear rate affect RBC dynamics, cell-free layer (CFL) thickness, and platelet margination near a sinusoidal wall. The sinusoidal wall, with crests and valleys aligned with the flow direction, serves as a model of the flow-aligned platelet aggregates observed in microfluidic experiments [Pero et al., CRPS, 2024]. We perform three-dimensional immersed-boundary-lattice-Boltzmann simulations of particulate blood flow with deformable RBCs and nearly rigid spherical platelets. Our results show that platelet margination is primarily governed by Ht and is more pronounced in regions where the CFL thickness is similar to the platelet size. At low Ht, platelets preferentially accumulate at crests, promoting high-amplitude aggregate growth. Increasing Ht leads to a more uniform platelet distribution along the surface, consistent with experimental observations. The sinusoidal geometry generates a pronounced crest-valley wall shear rate gradient, suggesting that distinct shear-dependent adhesion pathways may dominate at different surface locations. Our findings provide mechanistic insights into the morphological evolution of platelet aggregates and may ultimately inform targeted therapeutic strategies for thrombosis based on shear-sensitive drug-delivery.

The activity and distribution of Piezo1 molecules, along with the maturity and strength of focal adhesions (FAs), serve as critical factors influencing cell mechanosensing. Notably, migrating epithelial cells and mesenchymal-like cancer cells exhibit significantly different behaviors regarding these elements. In cancer cells, Piezo1 molecules are distributed uniformly, while in epithelial cells, their distribution is heterogeneous. In epithelial cells, Piezo1 molecules tend to group around FAs, a phenomenon that is enhanced by actomyosin contractility. However, a reduction in contractility results in a more uniform distribution of Piezo1 molecules. The expression and activity levels of Piezo1 molecules are markedly higher in cancer cells compared to epithelial cells. The activity of Piezo1 molecules correlates with the intracellular calcium concentration. Despite the extensive experimental studies on the properties of migrating epithelial and mesenchymal-like cancer cells, the physical explanations remain lacking. The primary objective of this theoretical study is to explore: (i) the inhomogeneous distribution of Piezo1 molecules in epithelial cells in relation to the Marangoni effect, (ii) the heightened activity of Piezo1 molecules in cancer cells by specifying the driving force, and (iii) the influence of membrane-mediated interactions among Piezo1 molecules grouped near FAs in epithelial cells on their activity.

Climate change is reshaping species interactions and movement across fragmented landscapes. Despite this, most mathematical models assume random diffusion, overlooking the influence of directed movement. Here, we develop a graph based reaction-diffusion-advection framework explicitly incorporating directional movement induced by environmental gradients. Our results show while diffusion promotes overall population persistence across the network, advective movement induces asymmetric flows. It create population hotspots by directing individuals toward optimal niches, often associated with nodes of high in-degree. We demonstrate the interplay between advection strength and network topology in determining species persistence. Strong advection increase local extinction risk by accumulating populations toward favorable nodes. Additionally, loss of ecological corridors can disrupt directed flow within the network, thereby restricting species from favorable patches. We found that this disruption might not cause immediate extinction, rather forcing species to spread to the suboptimal patches. Our advection framework therefore efficiently captures how directional movement interacting with network topology governs species redistribution, hotspot formation, and predict extinction risk under environmental change.

Representation learning on large-scale unstructured volumetric and surface meshes poses significant challenges in neuroimaging, especially when models must incorporate diverse vertex-level morphometric descriptors, such as cortical thickness, curvature, sulcal depth, and myelin content, which carry subtle disease-related signals. Current approaches either ignore these clinically informative features or support only a single mesh topology, restricting their use across imaging pipelines. We introduce a hierarchical transformer framework designed for heterogeneous mesh analysis that operates on spatially adaptive tree partitions constructed from simplicial complexes of arbitrary order. This design accommodates both volumetric and surface discretizations within a single architecture, enabling efficient multi-scale attention without topology-specific modifications. A feature projection module maps variable-length per-vertex clinical descriptors into the spatial hierarchy, separating geometric structure from feature dimensionality and allowing seamless integration of different neuroimaging feature sets. Self-supervised pretraining via masked reconstruction of both coordinates and morphometric channels on large unlabeled cohorts yields a transferable encoder backbone applicable to diverse downstream tasks and mesh modalities. We validate our approach on Alzheimer's disease classification and amyloid burden prediction using volumetric brain meshes from ADNI, as well as focal cortical dysplasia detection on cortical surface meshes from the MELD dataset, achieving state-of-the-art results across all benchmarks.

We introduce Mean-Field Game (MFG) epidemiological models, in which immunity either wanes with time in a fully observable way or disappears instantaneously with no direct observation (making a previously recovered individual fully susceptible again without realizing it). Both interpretations create computational challenges for rational noninfected individuals deciding on their contact rates based on their personal current immunity state and the changing epidemiological situation. Both require solving a forward-backward MFG system that includes PDEs (an advection-reaction equation for the immunity-structured population and a Hamilton-Jacobi-Bellman equation for the corresponding value function). We show how this can be done efficiently by solving a two-point boundary value problem for a system of approximating ODEs. We also show how the same approach can be extended to handle an initial uncertainty in the planning horizon.

We introduce eigencone constellations, a hierarchical framework for embedding bounded-degree spatial graphs into concentric spherical shells and partitioning each shell into spectrally weighted, spherical star-shaped territories. Given a connected sparse spatial graph $G$ with a distinguished root vertex (the queen), we assign each vertex to a sphere whose radial position is determined by its graph distance from the queen, then tessellate each sphere into constellation territories whose solid angles are proportional to the spectral mass of the corresponding subgraph. Within each territory, nodes are packed by constrained repulsion, yielding local simplex structures. The resulting geometric representation provides a structural framework for measuring spectral distance between dynamic subgraph states. By combining this eigencone-derived metric with constraints on the domain-specific edit alphabet, we define a forward-only deterministic trajectory -- the isomorphic walk -- which converges graph edits efficiently. We define the notion of spherical star-shaped domains with geodesic visibility, establish their properties under spectral projection, and demonstrate the trajectory convergence on molecular contact graphs.

Body Mass Index (BMI) is a widely accessible but imprecise proxy of cardiometabolic health. While assessing true body composition is superior, gold-standard methods like Dual-Energy X-ray Absorptiometry (DXA) are not scalable. We address this gap by developing and validating "PhotoScan," a method to estimate body composition from smartphone imagery. We pretrained a deep learning model on UK Biobank participants (N=35,323) and fine-tuned on a newly recruited clinical cohort (PhotoBIA cohort, N=677) with diverse ethnicity, age, and body fat distribution, achieving high accuracy against DXA for total body fat percentage (BF%, MAE = 2.15%), Android-to-Gynoid fat ratio (A/G, MAE = 0.11), and visceral-to-subcutaneous fat area ratio (V/S, MAE = 0.09). Generalizability of the model was demonstrated on an independent metabolic health study cohort (MetabolicMosaic cohort, N=132 participants), achieving MAEs of 2.13% for BF%, 0.09 for A/G, and 0.09 for V/S. We then evaluated the clinical utility of these metrics in the MetabolicMosaic cohort by predicting insulin resistance (IR). Adding PhotoScan-derived body composition metrics to baseline demographics model (Age, Sex, BMI) significantly improved insulin resistance classification (Area Under the Receiver Operating Characteristic Curve "AUROC" 76.0% vs 69.2%, DeLong test p=0.002, Net Reclassification Index "NRI" 0.593). Crucially, this accessible smartphone method achieved performance nearly equivalent to adding clinical-grade DXA data to baseline demographics model (AUROC 77.3% vs 69.2%, DeLong test p=0.004, NRI 0.748). These findings demonstrate that smartphone-based phenotyping captures clinically meaningful risk signals missed by BMI and anthropometrics, offering a scalable alternative to DXA for cardiometabolic risk stratification.

Human eye-hand coordination relies on internal forward models that predict future states and compensate for sensory delays. During line tracing, the gaze typically leads the hand through predictive saccades, yet the extent to which this predictive window reflects expertise or intrinsic individual traits remains unclear. In this study, I examined eye-hand coordination in professional calligraphers and non-experts performing a controlled line tracing task. The temporal coupling between saccade distance (SD) and pen speed (PS) revealed substantial interpersonal variability: SD-PS peak times ranged from approximately -50 to 400 ms, forming stable, participant-specific predictive windows that were consistent across trials. These predictive windows closely matched each individual's pen catch-up time, indicating that the oculomotor system stabilizes fixation in anticipation of the hand's future velocity rather than relying on reactive pursuit. Neither the spatial indices (mean gaze-pen distance, mean saccade distance) nor the temporal index (SD-PS peak time) differed between calligraphers and non-calligraphers, and none of these predictive parameters correlated with tracing accuracy. These findings suggest that diverse predictive strategies can achieve equivalent performance, consistent with the minimum intervention principle of optimal feedback control. Together, the results indicate that predictive timing in eye-hand coordination reflects a stable, idiosyncratic Predictive Protocol shaped by individual neuromotor constraints rather than by expertise or training history.

Advances in AI have introduced several strong models in computational pathology to usher it into the era of multi-modal diagnosis, analysis, and interpretation. However, the current pathology-specific visual language models still lack capacities in making the diagnosis with rigorous reasoning paths as well as handling divergent tasks, and thus, challenges of building AI Copilots for real scenarios still exist. Here we introduce TeamPath, an AI system powered by reinforcement learning and router-enhanced solutions based on large-scale histopathology multimodal datasets, to work as a virtual assistant for expert-level disease diagnosis, patch-level information summarization, and cross-modality generation to integrate transcriptomic information for clinical usage. We also collaborate with pathologists from Yale School of Medicine to demonstrate that TeamPath can assist them in working more efficiently by identifying and correcting expert conclusions and reasoning paths. We also discuss the human evaluation results to support the reasoning quality from TeamPath. Overall, TeamPath can flexibly choose the best settings according to the needs, and serve as an innovative and reliable system for information communication across different modalities and experts.

Biological time can be measured in two ways: in generations and in physical (chronological) time. When generations overlap, these two notions diverge, which impedes our ability to relate mathematical models to real populations. In this paper we show that nevertheless, the two clocks can be synchronised in the long run via a simple identity relating generational and physical time. This equivalence allows us to directly translate statements from the generational picture to the physical picture and vice versa. We derive a generalized Euler-Lotka equation linking the basic reproduction number $R_0$ to the growth rate, and present a simple identity that relates the selection coefficient of a mutation to the history of typical individuals, with applications to epidemiology, population biology and microbial growth.

How does the human brain encode complex visual information? While previous research has characterized individual dimensions of visual representation in cortex, we still lack a comprehensive understanding of how visual information is organized across the full range of neural population activity. Here, analyzing fMRI responses to natural scenes across multiple individuals, we discover that neural representations in human visual cortex follow a remarkably consistent scale-free organization -- their variance systematically decays as a power law, detected across four orders of magnitude of latent dimensions. This scale-free structure appears consistently across multiple visual regions and across individuals, suggesting it reflects a fundamental organizing principle of visual processing. Critically, when we align neural responses across individuals using hyperalignment, we find that these representational dimensions are largely shared between people, revealing a universal high-dimensional spectrum of visual information that emerges despite individual differences in brain anatomy and visual experience. Traditional analysis approaches in cognitive neuroscience have focused primarily on a small number of high-variance dimensions, potentially missing crucial aspects of visual representation. Our results demonstrate that visual information is distributed across the full dimensionality of cortical activity in a systematic way, suggesting we need to move beyond low-dimensional characterizations to fully understand how the brain represents the visual world. This work reveals a new fundamental principle of neural coding in human visual cortex and highlights the importance of examining neural representations across their full dimensionality.