2026-07-25 每日论文

来源	独立页面
arXiv	arXiv
bioRxiv	bioRxiv
medRxiv	medRxiv

arXiv

抓取失败：Page request resulted in HTTP 503 (https://export.arxiv.org/api/query?search_query=cat%3Acs.NE+OR+cat%3Acs.MA+OR+cat%3Acs.LG+OR+cat%3Acs.CV+OR+cat%3Acs.CL+OR+cat%3Acs.AI+OR+cat%3Astat.ML+OR+cat%3Aq-bio.BM+OR+cat%3Aq-bio.CB+OR+cat%3Aq-bio.GN+OR+cat%3Aq-bio.MN+OR+all%3A%22large+language+model%22+OR+all%3A%22large+language+models%22+OR+all%3ALLM+OR+all%3ALLMs+OR+all%3A%22foundation+model%22+OR+all%3A%22foundation+models%22+OR+all%3A%22generative+AI%22+OR+all%3A%22mixture+of+experts%22+OR+all%3A%...

bioRxiv

标题	作者	发布日期	PDF链接	摘要
Global structure of new relational knowledge networks is represented in retrosplenial complex, and node-distance in hippocampus	Feld, G. B.	2026-07-25	PDF	Much of our knowledge is structured in networks, but how the brain represents such networks remains unclear. In most nested knowledge structures a subset of the nodes will be highly connected to other nodes in the network. High levels of connection can exist either locally (high degree centrality) or at a global level (high closeness/betweenness centrality). Here, we explored how the human brain represents network structure using functional magnetic resonance imaging (fMRI) and a novel learning paradigm. During training, participants learned the transition structure for travel between a set of fictious partially connected alien planets. Despite encountering only individual connections, participants choices indicated they could infer the broader network structure. Next day, they viewed each stimulus during a cover task while undergoing fMRI. Representational-similarity analysis showed that shortest-path distances within the learned graph were encoded in the posterior hippocampus and right retrosplenial complex, whereas global connectivity (closeness/ betweenness centrality) was represented in the left retrosplenial complex. These findings extend the view that brain networks associated with spatial navigation also process abstract relational knowledge using principles akin to mapping physical space. The results are also consistent with proposals that the hippocampus represents distance information and that the retrosplenial cortex acts as hub for integrating recently acquired knowledge.
Deep Phenotyping with Global Brain Activity and Plasticity Mapping Identify the Dorsal Raphe-Basolateral Amygdala Circuit as a Mediator of Adaptive Stress Responses	Mitra, S.	2026-07-25	PDF	Exposure to chronic environmental challenges triggers divergent behavioral trajectories across individuals. At the core, these different trajectories can be classified as individuals actively adapting to the challenges (responders) and those displaying a rigid, non-responsive phenotype (non-responders). The brain system-wide network configurations that dictate why individuals diverge along these differential coping strategies, which can also lead to disease vulnerability or resilience, remain poorly understood. Here, we paired machine-learning-based deep behavioral phenotyping with multi-modal whole-brain imaging, integrating longitudinal Manganese-Enhanced MRI (MEMRI) and post-challenge cFOS mapping, to chart the functional landscape of individual stress trajectories in mice subjected to chronic social defeat stress. High-dimensional behavioral phenotyping revealed that active stress adaptation is a complex trajectory marked by latent, pre-stress kinetic signatures in vigilance-like and locomotive behaviors. At the neural level, longitudinal MEMRI captured distinct, consolidated activity reconfigurations across canonical valence and stress-regulatory circuits that segregated responders from non-responders. Complementary whole-brain cellular cFOS network analysis after an additional acute challenge revealed that non-responders exhibited marked hyper-modularity and network fragmentation, whereas responders featured a tightly integrated functional module co-clustering the periaqueductal gray, ventral tegmental area, basolateral amygdala (BLA), and dorsal raphe (DR). Notably, functional network connectivity along the DR-BLA axis was completely lost in non-responsive animals. Finally, pathway-specific chemogenetic inhibition of BLA-projecting DR neurons during a social challenge significantly attenuated social avoidance and reversed anxiety-like behavioral deficits, effectively shifting active behavioral adaptation toward a non-responsive phenotype. Together, these findings demonstrate that individual stress-coping strategies are driven by coordinated, system-wide reconfigurations of activity and plasticity, identifying the DR-BLA circuit as a critical gatekeeper of adaptive stress responses.
A significant enrichment that is not: spatial nulls, co-expression, and the imaging transcriptomics of EEG alpha-power genetics	Schenetti, J.	2026-07-25	PDF	Background. Imaging transcriptomics routinely asks whether a trait-associated gene set is over-expressed in a region of interest, and the field standard is to guard that inference with a spatial-autocorrelation-preserving spin test. Electroencephalographic (EEG) oscillatory power is among the most heritable human neurophysiological traits, and the cortical generators of the alpha rhythm have been characterised independently from resting-state magnetoencephalography - making this a natural test bed both for asking whether trait genetics is regionally organised, and for asking what such a test actually establishes. Objective. To test whether alpha-associated genetic signal is spatially enriched in the cortical generators of the alpha rhythm, and to evaluate that inference against complementary null models. Methods. MAGMA gene-based analysis of ENIGMA-EEG summary statistics for six phenotypes: central and occipital alpha power, occipital alpha peak frequency, and theta, beta and delta power. Regional transcription was obtained from the Allen Human Brain Atlas (AHBA) with abagen in the Glasser HCP-MMP1.0 atlas, with Schaefer-100 and Yan-600 as sensitivity analyses. Enrichment in the 41 cortical alpha-source regions was quantified with a threshold-free continuous score and a top-100 gene-set composite, and assessed against three complementary nulls: a spin test (10,000 rotations, cross-checked against brainsmash surrogates), a co-expression-aware gene-set null (10,000 matched random gene sets), and a positive control on a known expression gradient. Results. Judged by the field-standard spin test alone, this study would have reported a positive, biologically coherent finding: alpha-power genes are enriched in the cortical alpha generators (continuous p_spin = 0.022; top-100 p_spin = 0.030), with the spin result corroborated by an independent surrogate model (p = 0.018) and the pipeline validated by a positive control (p_spin = 2 x 10^-4). Three further tests dissolve that conclusion. First, it is not band-specific: theta, beta and delta enrich comparably or more strongly (top-100 beta p_spin = 0.011; delta 0.042; continuous theta 0.042). Second, it does not replicate across alpha phenotypes (occipital alpha continuous p_spin = 0.20; alpha peak frequency non-significant, p_spin >= 0.066). Third, against random gene sets of matched size the alpha set is unremarkable (p_geneset = 0.33) - the apparent enrichment is a generic property of arbitrary gene sets in this cortical territory, and is invisible to a spatial-only null. No test survived false-discovery-rate correction across the 24-cell phenotype x score x region-set grid (minimum q = 0.127), and nominal significance did not survive a change of parcellation. Conclusion. EEG alpha-power genetics shows no regionally specific transcriptomic signature in the cortical generators of the rhythm; the weak tendency that is present is shared across frequency bands, consistent with their known genetic correlation. Methodologically, this is a worked demonstration that correcting for spatial autocorrelation is necessary but not sufficient: a spin-significant, surrogate-corroborated, mechanistically plausible enrichment can be fully accounted for by gene-set co-expression. Enrichment claims in imaging transcriptomics should report a gene-set null alongside the spatial null.
Distinct forms of dopamine transmission control locomotion and learning	McGregor, M.	2026-07-25	PDF	The neuromodulator dopamine is essential for voluntary movement and learning from experience. While these behaviors often occur simultaneously and rely on overlapping nigrostriatal dopamine circuits, they are also separable and can manifest independently. How a single neuromodulatory system controls such disparate aspects of behavior in parallel remains poorly understood. Here we show that striatal dopamine is released through two spatially and functionally distinct modes that are independently regulated and serve distinct behavioral roles. Eliminating dopamine release underlying bulk extracellular accumulation reveals a spatially restricted mode of transmission that maintains phasic dopamine receptor activation on striatal projection neurons. We find that selective loss of diffuse dopamine impairs striatal circuit excitability and reduces locomotion, while point-to-point transmission is sufficient to maintain striatal spine density and support motor and associative learning. We propose that the geometry of dopamine release determines its behavioral output, a principle that explains the breadth of dopamine's functions and more broadly elucidates how neurons multiplex different forms of information.
Live-imaging of endogenous neurofascins reveals glial adhesion shapes developing nodes of Ranvier	Lyster-Binns, P.	2026-07-25	PDF	The organisation of myelinated axons into specialised domains is essential for saltatory conduction and depends on the polarised distribution of neuronal and glial cell-adhesion molecules, including neurofascins. However, how these domains assemble and refine in vivo remains unclear, as longitudinal imaging of endogenous proteins within intact nervous systems remains challenging. Here, we generated knock-in zebrafish in which endogenous neuronal and glial neurofascins are fused to fluorescent proteins, which enabled live-imaging without perturbing node assembly or introducing overexpression artefacts. Using these reporters, we visualised neuronal and glial neurofascin dynamics, and the formation of nodes of Ranvier and paranodes in vivo. We find that nascent nodes progressively compact in developing peripheral nerves, and that this process depends on glial neurofascin. Our novel toolkit for imaging endogenous neurofascins reveals a glia-dependent mechanism that shapes the developmental refinement of nodal architecture after their initial assembly. Our findings imply that glia-driven modulation of paranodes can be a mechanism for nervous systems to regulate circuit function.
Developmental tuning of functional manifold dimensionality across the human brain	Busch, E. L.	2026-07-25	PDF	Neural representations vary in their complexity across brain regions and tasks. How this variation emerges over human development remains poorly understood. We estimated intrinsic dimensionality in five naturalistic fMRI datasets (N = 781 unique participants, aged 3 months to 53 years) with T-PHATE -- a nonlinear manifold learning method robust to noisy, autocorrelated signals. In adults, brain regions relevant to a task had higher-dimensional activity than task-irrelevant regions across auditory, visual, and audiovisual stimuli. The modulation of representational complexity by tasks was absent in infants, emerged in early childhood, and strengthened logarithmically through adolescence. It reflected a selective collapse in dimensionality in task irrelevant regions, relative to a resting-state baseline, rather than an expansion of dimensionality in task-relevant regions. Such compression followed a trajectory from global and nonselective in infants to local and precise by adulthood. These results identify selective compression as a developmental engine of functional specialization: rather than adding complexity where it is needed, the brain dynamically pares it away where it is not.
AI segmentation requires accounting for brain size to maintain performance on developmental MRI cohorts	Dorfschmidt, L.	2026-07-25	PDF	The human brain undergoes rapid developmental changes through early life, underpinning the emergence of function but also marking a period of vulnerability to a range of neurodevelopmental disorders. With dynamic changes to brain size, morphology, and imaging contrast, consistent and accurate computational neuroanatomy remains a challenge. Deep learning tools for segmentation, like SynthSeg, offer robustness to heterogeneously acquired MRI contrast but remain unproven in early development. Here, we aggregated a large cohort (26k) of MRI scans spanning infant to adult development, and evaluated SynthSeg performance. Automated quality control scores, visual inspection, and spatial overlap with expert-segmented MRI scans revealed poor quality output segmentations during development. In the infant period only 36% of scans (1094/3069) passed automated QC. Rescaling infant scans to adult brain sizes significantly improved spatial overlap, and cropping scans to match adult fields of view retrieved automated quality control. Evaluation of the SynthSeg rescale + crop pipeline demonstrated visible and quantitative improvements in segmentation throughout infancy and childhood. There were marked increases in successful segmentations in infant scans, with 91% of scans now passing QC (2803/3069). These findings facilitate computational analysis of typical and disrupted neurodevelopment and should be considered when training the next generation of computational tools.
Humans could become the greatest driver of biosphere net gain in Earth history, but we are currently the second fastest driver of biosphere loss	Wong Hearing, T. W.	2026-07-25	PDF	Human activity is transforming the shape, size, and resilience of Earths biosphere, degrading and augmenting Holocene baseline conditions at various scales, and replacing the wild biosphere with an anthropogenically modified one. We evaluate episodes of biosphere change throughout Earth history and compare them with contemporary and near-future anthropogenic changes, developing the concept of biosphere disruptors - agents that force global-scale macroevolutionary change. Transient disruptors are short-lived agents (mean 8.0x105 years), including massive volcanism and asteroid impacts. Persistent disruptors, including atmospheric and ocean oxygenation and land plant evolution, remain in the Earth System over long timescales (mean 1.6x108 years). In the geological record, transient disruptors are associated with temporary but sometimes massive biosphere degradation, whereas persistent disruptors are associated with sustained biosphere enhancement. Most anthropogenic biosphere impacts resemble those of past transient disruptors, globally degrading wild biomass and biodiversity. Humanity is driving the second highest rate of biosphere degradation in Earth history after the Cretaceous-Palaeogene bolide impact. However, humanity is the first disrupting agent capable of reflecting on and potentially transforming its impact on planetary habitability. If we can achieve this, humanity could drive the greatest rate of increase in planetary habitability in Earth history on centennial to millennial timescales.
Generative replay across hippocampal-neocortical circuits	Shearer, C. M.	2026-07-25	PDF	To make flexible decisions, the brain needs to make inferences between events that were not directly experienced together. In sharp-wave ripple (SWR) events, during periods of rest and sleep, spiking activity in the hippocampus appears to support this process, by co-activating mnemonic representations of discrete cues and events that have not been experienced together. However, it remains unclear whether this 'generative replay' also occurs in other brain regions to update beliefs across the brain. To address this question, here we use multi-unit electrophysiology and calcium imaging in freely moving mice during a multi-day inference task. We show evidence for a putative anatomical pathway from the dorsal CA1 (dCA1) region of the hippocampus to primary visual cortex (V1), via the granular Retrosplenial Cortex (RSCg). We then characterise the neuronal activity across this pathway during an inference task and in subsequent periods of sleep. We show that during inference, mnemonic activity in V1 mirrors that observed in dCA1 to represent learned associations. In sleep, we show evidence for generative replay in V1, where co-activity can be observed between cells representing inferred relationships between cues that have not been directly experienced together. This activity in V1 is predicted by activity in dCA1, suggesting that the hippocampus coordinates generative replay across neocortical circuits. This coordinated generative replay may underpin the formation of a hierarchical generative model capable of predicting future scenarios that extend beyond direct experience.
Concept-Level Semantic Representations Remain Decodable in Chronic Post-Stroke Aphasia	Swiderski, A. M.	2026-07-25	PDF	Aphasia is characterized by impaired word retrieval, yet most cognitive models of word production assume that underlying conceptual-semantic representations are largely preserved. This study investigated whether concept-level semantic structure remains decodable from BOLD signals in chronic post-stroke aphasia and which semantic models best explain neural representational geometry during covert semantic feature generation. Eight healthy adults and six individuals with chronic aphasia completed a dense-sampling fMRI protocol in which they viewed 57 pictured nouns while silently generating semantic features. Representational similarity analysis showed that an experiential model (Exp48) best matched neural geometry in both people with aphasia and controls, outperforming taxonomic (WordNet) and distributional (Word2Vec, GloVe) models. Using representational similarity decoding, concept identity was recovered well above chance in both groups. No relationship was found between decoding accuracy and language measures from individuals with aphasia. These findings suggest that experiential semantic structure remains robustly represented and decodable in chronic aphasia despite lesion-related language impairments, highlighting preserved conceptual representations alongside altered anatomy.

medRxiv

标题	作者	发布日期	PDF链接	摘要
A Machine Learning Based Causal Interface for Time-Varying Environmental Predictors of Substance Use Initiation in the ABCD Study	Wei, M.	2026-07-25	PDF	Background: The Adolescent Brain Cognitive Development (ABCD) Study offers rich longitudinal data on environmental, genetic, and other factors related to substance use initiation. Classical marginal structural models (MSMs) require choosing which covariates to include in propensity models, but this choice is difficult in the presence of hundreds of correlated predictors. Methods: We analyzed longitudinal data from the Adolescent Brain Cognitive Development Study using a European-ancestry unrelated cohort, where each individual contributed repeated observations over time. Interval level binary outcomes were defined for initiation of alcohol, nicotine, cannabis, and any substance, restricting analyses to participants at risk prior to initiation. All predictors were constructed as lagged variables to preserve temporal ordering. We implemented a two-stage machine learning based causal framework. First, we performed graph discovery using a Granger-inspired lagged predictive modeling approach, applying elastic-net logistic regression to identify predictive relationships between lagged environmental variables and future initiation outcomes. Robust candidate edges were selected using subject level bootstrap stability selection. Second, we estimated adjusted effect sizes for stable edges using DML-style partialling-out with cross-fitting. For each candidate predictor, we defined the treatment as the lagged variable of interest and adjusted for high dimensional lagged covariates. Cross fitting with group-based splitting accounted for within-subject dependence, and nuisance functions were estimated using random forest models. Cluster robust standard errors were used for inference. Results: Across analyses, behavioral and environmental predictors showed stronger and more consistent associations with substance use initiation than polygenic risk scores. In the graph discovery analysis, stable edges were mainly related to rule-breaking behavior, sensation seeking, resiliency, sleep, screen related measures, and family or parent related variables. PRS variables were less consistently selected and did not appear as dominant predictors. Cannabis initiation showed strong stable links with parent-reported rule breaking behavior and behavioral symptoms. Nicotine initiation was linked to sensation seeking, behavioral symptoms, and screen-related measures. DML style effect estimates were modest in magnitude but supported positive associations for several stable predictors, including rule breaking behavior for cannabis initiation and sensation seeking, BPM behavioral items, and screen-related measures for nicotine initiation. Conclusions: In this EUR unrelated cohort, adolescent substance-use initiation was more consistently associated with modifiable behavioral and environmental factors than with PRS variables. The findings suggest that cannabis and nicotine initiation share some risk factors but also have distinct predictor profiles. These results highlight the importance of longitudinal, multi method approaches for identifying early risk patterns and suggest that prevention strategies may benefit from focusing on behavioral regulation, sensation seeking, family context, sleep, and screen-related behaviors.
Dynamic and Baseline Multi-Task Learning for Predicting Substance Use Initiation in the ABCD Study	Wei, M.	2026-07-25	PDF	Background: Substance use initiation during adolescence is a time-dependent process influenced by shared and substance-specific risk factors. Conventional single-outcome models may not capture nonlinear relationships, shared structure across outcomes, or changes in risk over time. We compared baseline and dynamic multi-task learning frameworks for predicting initiation of alcohol, nicotine, cannabis, and any substance use. Methods: We analyzed 2,366 unrelated participants of European genetic ancestry from the Adolescent Brain Cognitive Development Study release 5.1. A baseline multi-task learning model predicted initiation within 48 months using one record per participant. A dynamic discrete-time multi-task learning model used longitudinal interval records to estimate time-varying initiation risk. Both frameworks included environmental and behavioral exposures, core covariates, and polygenic risk scores. Performance was evaluated in a held-out test set using area under the receiver operating characteristic curve, precision-recall area under the curve, and calibration metrics. Multi-task models were compared with corresponding single-task logistic regression models. Permutation-based feature importance was compared across baseline and dynamic models and with Cox proportional hazards results. Results: Initiation rates were 40.7% for alcohol, 6.4% for nicotine, 4.2% for cannabis, and 43.4% for any substance use. Multi-task learning did not consistently outperform logistic regression across all outcomes. Its clearest gains were observed for the lower-prevalence nicotine and cannabis outcomes. Dynamic modeling generally improved prediction relative to baseline modeling, particularly for logistic regression and nicotine initiation, whereas baseline multi-task learning remained stronger for cannabis. Feature rankings showed moderate agreement between baseline and dynamic multi-task models but lower concordance between multi-task and Cox models. Behavioral and environmental predictors, particularly UPPS sensation seeking and parental monitoring, were more reproducible across frameworks than polygenic risk score features. Conclusions: Longitudinal discrete-time modeling provided the most consistent improvement in predicting adolescent substance use initiation, while the added value of multi-task learning varied by outcome and was most apparent for nicotine and cannabis. Combining baseline and dynamic analyses may help identify robust shared and substance-specific predictors while characterizing how risk changes over development.
Circulating protein profiling identifies prognostic biomarkers in amyotrophic lateral Sclerosis	Klimovski, H.	2026-07-25	PDF	In the pathologically and clinically heterogeneous neurodegenerative disorder amyotrophic lateral sclerosis (ALS), objective biochemical predictors of survival are essential to handle complexity in clinical trials, enrich clinical decision-making and interrogate the biology of disease progression. In this longitudinal study, we performed high-depth proximity extension assay proteomics using 1,095 samples of serum (N=851) and CSF (N=244) from 426 people with ALS, with orthogonal replication in an external cohort of 349 people with ALS. Age- and sex-adjusted Cox analysis identified 57 proteins in serum, including neurofilament light chain (NEFL) and peripherin, as well as five proteins in CSF, including tropomyosin 3 (TPM3) that were associated with survival (FDR-adjusted p<0.05). Penalised Cox regression identified a panel of 9 serum proteins - including NEFL, peripherin, TNF receptor superfamily member 27 (EDA2R) and calcitonin - that reflect the extent of disease as well as the progression rate, improving survival prediction compared with models using clinical parameters and NEFL. Joint modelling identified associations between the longitudinal trajectories of serum EDA2R and calcitonin with survival, highlighting their potential role in measuring disease progression. This work indicates the utility of multiple proteins reflecting diverse biological pathways in refining survival stratification and highlights systemic factors in ALS progression.
CT-Based Deep Foundation Model for Predicting Immune Checkpoint Inhibitor-Induced Pneumonitis Risk in Lung Cancer	Muneer, A.	2026-07-25	PDF	Background: Immune checkpoint inhibitors (ICIs) have revolutionized cancer therapy but can cause serious immune-related adverse events (irAEs), with pneumonitis (ICI-P) being among the most severe. Early identification of high-risk patients before ICI initiation is critical for close monitoring, timely intervention, and optimizing outcomes. Purpose: To develop and validate a deep learning foundation model to predict ICI-P from baseline CT scans in patients with lung cancer. Methods: We designed the Checkpoint-Inhibitor Pneumonitis Hazard EstimatoR (CIPHER), a deep learning-powered foundation model combining contrastive learning with a transformer-based masked autoencoder to predict ICI-P from baseline CT scans in lung cancer patients. Using self-supervised learning, CIPHER was pretrained on 590,284 CT slices from 2,500 non-small cell lung cancer (NSCLC) patients to learn representations of heterogeneous lung parenchyma. Following pretraining, CIPHER was adapted to the internal MDA NSCLC immunotherapy cohort of 347 patients, of whom 33 developed adjudicated ICI-P. Fine-tuning was performed using 254 non-ICI-P patients only, and a held-out internal validation set of 93 patients, including 33 ICI-P cases and 60 non-ICI-P controls, was reserved for evaluation. CIPHER was benchmarked against clinical, radiomics, and ensemble comparator models and externally validated in an independent Johns Hopkins NSCLC cohort of 116 patients, including 20 ICI-P cases and 96 non-ICI-P controls. Results: In our internal immunotherapy cohort, CIPHER consistently distinguished patients at elevated risk of ICI-P from those without the event, with AUCs ranging from 0.77 to 0.85. In head-to-head benchmarking, CIPHER achieved an AUC of 0.83, compared with 0.58 for the clinical model, 0.77 for the radiomics model, and 0.78 for the ensemble model. In the external validation cohort, CIPHER maintained high performance, AUC = 0.83 and balanced accuracy = 81.7%, exceeding the radiomics model, DeLong p = 0.0318, and showing substantially higher specificity than the radiomics model, 83.3% versus 45.8%, with sensitivity of 80.0%. Confusion matrix analyses showed that CIPHER correctly identified 80 of 96 non-ICI-P cases and 16 of 20 ICI-P cases. Conclusions: We developed and externally validated CIPHER, a CT-based imaging biomarker for pretreatment ICI-P risk stratification in NSCLC. CIPHER shows promise as a noninvasive tool for ICI-P risk assessment but warrants prospective validation before clinical translation.
Why More Doctors May Not Mean More Essential-Specialty Physicians	Yu, H.	2026-07-25	PDF	BackgroundI present an illustrative conceptual model of how economic and legal disincentives may drive specialty and practice-setting misallocation in South Korean essential medicine, alongside--not instead of--aggregate workforce constraints, examined through the "Vital 5" specialties (Internal Medicine, General Surgery, Obstetrics & Gynecology, Pediatrics, and Cardiothoracic Surgery). MethodsI built two models: a Human-Capital Misallocation Cost model estimating resources lost when specialists work outside their training, and an Expected Value (E(V)) model of entry into essential fields incorporating Net Return (Rnet), Probability of Lawsuit during a practicing physicians lifetime (Psuit), and Cost of Risk (Crisk). All parameters are stated assumptions with sensitivity analyses. A national replacement-cost equivalent was derived from emergency-system expenditure per resident vacancy in the 2024 crisis and a mismatch rate taken directly from national data. ResultsNational data show a high proportion of cardiothoracic surgeons practicing outside their major field. Under the stated assumptions, expanding admissions dilutes Rnet and, through undertraining, raises Psuit; E(V) then declines and falls below zero only when litigation exposure is assumed to rise, remaining positive when it is held constant. The reversal therefore depends on the assumed expansion-litigation link, not on income dilution alone. ConclusionI propose, as a model-derived hypothesis, a "Paradox of Expansion": expanding quotas without addressing compensation and legal risk may reduce, rather than increase, active specialists in selected essential fields. I compare no alternative policies and identify no optimal policy; I argue only that quota expansion alone is unlikely to improve essential-specialty retention unless reimbursement and medico-legal risk are addressed concurrently.
Cardiovascular-Kidney-Metabolic Health in US Adults Under the 2026 Multisociety Guideline: Stage Redistribution From 1999 to 2023 and Population Burden Through 2050	Fu, F.	2026-07-25	PDF	Background The 2026 multisociety guideline frames cardiovascular, kidney, and metabolic health as a staged pathway. National surveillance should distinguish early-risk expansion from advanced-disease accumulation because they require different responses. Methods We analysed 66,553 adults aged 20 years or older across 11 non-overlapping National Health and Nutrition Examination Survey periods from 1999-2000 to August 2021-August 2023. Staging incorporated adiposity, glycaemia, metabolic risk, KDIGO kidney risk, PREVENT 10-year cardiovascular disease risk, and clinical cardiovascular disease. We estimated age-standardised prevalence and secular trends and projected population burden through 2050. Robustness analyses included complete-case estimation, alternative staging thresholds, exclusion of the latest survey period, and rolling temporal validation. Findings In 2021-2023, 88.5% (95% CI 87.2-89.9) of US adults were in stage 1 or higher, and 62.1% (59.9-64.4) were in stages 2-4. From 1999-2000 to 2021-2023, stage 0 decreased by 4.3 percentage points and stage 1 increased by 7.9 points, whereas stages 3-4 remained stable. Adiposity and diabetes increased, while hypertension and hypertriglyceridaemia declined. The prevalence of stages 2-4 ranged from 54.4% among college graduates to 69.6% among adults with less than a high-school education. Under population ageing alone, 189.5 million adults were projected to be in stages 2-4 by 2050. Rolling temporal validation yielded a mean absolute error of 1.6 percentage points. Interpretation The distribution of CKM stages in the US population shifted towards earlier metabolic risk without a parallel increase in advanced-stage disease. Policy should combine upstream prevention, equitable access to detection and treatment, and preparation for a growing older population requiring integrated care.
Circulating protein profiling identifies prognostic biomarkers in amyotrophic lateral Sclerosis	Klimovski, H.	2026-07-25	PDF	In the pathologically and clinically heterogeneous neurodegenerative disorder amyotrophic lateral sclerosis (ALS), objective biochemical predictors of survival are essential to handle complexity in clinical trials, enrich clinical decision-making and interrogate the biology of disease progression. In this longitudinal study, we performed high-depth proximity extension assay proteomics using 1,095 samples of serum (N=851) and CSF (N=244) from 426 people with ALS, with orthogonal replication in an external cohort of 349 people with ALS. Age- and sex-adjusted Cox analysis identified 57 proteins in serum, including neurofilament light chain (NEFL) and peripherin, as well as five proteins in CSF, including tropomyosin 3 (TPM3) that were associated with survival (FDR-adjusted p<0.05). Penalised Cox regression identified a panel of 9 serum proteins - including NEFL, peripherin, TNF receptor superfamily member 27 (EDA2R) and calcitonin - that reflect the extent of disease as well as the progression rate, improving survival prediction compared with models using clinical parameters and NEFL. Joint modelling identified associations between the longitudinal trajectories of serum EDA2R and calcitonin with survival, highlighting their potential role in measuring disease progression. This work indicates the utility of multiple proteins reflecting diverse biological pathways in refining survival stratification and highlights systemic factors in ALS progression.
Operationalising the WHO call for integrated emergency care through a national ambulance alliance network: implementation experience and lessons from Addis Ababa, Ethiopia (PRECOS-1)	Dula, P. K.	2026-07-25	PDF	Prehospital emergency care in many African cities is constrained not by ambulance scarcity but by fragmentation: multiple uncoordinated provider types operating parallel dispatch systems, with no shared data and no capacity to measure system performance. Despite the World Health Assembly's 2023 resolution on integrated emergency care (WHA 76.2) and Ethiopia's Health Sector Transformation Plan II (HSTP-II), prehospital coordination has remained a critical missing link. In response, Ethiopia developed the Hospital Emergency Assistance and Response Tracking System (HEARTS) under the Ethiopian Ambulance Alliance Network (EAAN), a governed, multi-provider federation established through consultation with the Network for Perioperative and Critical Care (N4PCc), Federal Ministry of Health, the Addis Ababa city administration fire and disaster response team, a humanitarian NGO provider, and dedicated private ambulance providers. HEARTS provides real-time fleet tracking, unified call processing, community access via a dedicated mobile application, and performance dashboards disaggregated by sub-city and provider type. This paper reports the first phase of the Prehospital Care Outcome Study (PRECOS-1), a programmatic research platform established to generate longitudinal evidence on prehospital coordination in Ethiopia and to inform replication across Africa. During the first operational phase in Addis Ababa (December 2025 to June 2026), HEARTS coordinated 1,403 trips across four provider types and 59 response units. The majority of trips were high- or critical-acuity (63.3%), with maternal and obstetric presentations forming the largest clinical category (44.3%). Median response time was 14.5 minutes (IQR 5.8 to 27.0; 90th percentile 75.0 minutes), the first unified prehospital performance baseline for this city. These findings demonstrate that a coordination-first approach to prehospital system strengthening is feasible in a fragmented low- and middle-income country setting, generate the measurement infrastructure for future improvement studies (PRECOS-2 onward), and offer a replicable model for African cities pursuing integrated emergency care.
Pre-migratory, Migratory and Post-migratory Factors Associated with Risky Sexual Behaviour among Women Informal Cross-Border Traders at the COMESA Market, Lusaka, Zambia	Mwale, S.	2026-07-25	PDF	Women engaged in informal cross-border trade in Southern Africa experience frequent mobility, economic pressure, and exposure to new social environments that may increase vulnerability to risky sexual behaviour and HIV. Limited evidence exists on how pre-migratory, migratory, and post-migratory factors interact to shape sexual risk among this population. This study examined the prevalence and determinants of risky sexual behaviour among women informal cross-border traders at COMESA Market in Lusaka, Zambia. A cross-sectional survey was conducted with 499 women informal cross-border traders. A composite risk score was used to classify respondents as "low risk" or "high risk" for sexual behaviour. Descriptive statistics, chi-square tests, and multivariate binary logistic regression were used to investigate associations between selected covariates and risky sexual behaviour, guided by migration and vulnerability theory. The majority (67.7%) of respondents were classified as high-risk. In multivariate analysis, women who used personal savings for start-up capital were 73% less likely to be high-risk compared to those relying on nuclear family capital (aOR = 0.27, 95% CI: 0.10-0.73). Women with 5-9 years of trading experience also had lower odds (aOR = 0.21, 95% CI: 0.29-0.96). Conversely, using Katima Mulilo border increased the odds 15-fold (aOR = 14.59, 95% CI: 1.26-169.50). Not feeling disconnected while away (aOR = 3.39, 95% CI: 1.03-11.11) and sharing accommodation with one roommate (aOR = 4.42, 95% CI: 1.10-17.84) were also significant predictors of high-risk behaviour. This study demonstrates that risk among women informal cross-border traders is produced across the migration continuum by economic dependence, specific transit routes, and destination living conditions. Interventions should therefore prioritise financial independence, targeted health and protection services at high-risk borders, and safe accommodation with peer-led HIV prevention to reduce vulnerability among mobile women traders.
Two to Tango: Are Spouses' Perception of Women's Empowerment Associated with Concordance on Fertility Desires?	Kumari, B.	2026-07-25	PDF	Introduction The current study examined the association between husbands' and wives' perceptions of women's empowerment, focusing on household decision-making, attitudes toward IPV, and fertility desires in Pakistan. Methods Data from currently married couples participating in the Pakistan DHS (2017-18) were included in the analysis (n=3,027). The level of concordance on fertility desires was assessed using Kappa Coefficient. Principal component analysis (PCA) was applied to assess the internal consistency and dimensionality of household decision-making and IPV attitude items; items within each construct were subsequently summed to generate categorical empowerment scores used in the regression analyses. Relative risk ratios were estimated for perceived empowerment on spousal concordance on fertility desires using multinomial logistic regressions. Results The study found a moderate level of spousal agreement on fertility desires (Kappa=0.51). Among significant findings: wife's perception of her own high decision-making authority was associated with more than a two-fold increase in wife-only discordance (wife wanting more children, husband wanting no more), compared to both spouses wanting more children. Conclusion The study contributes to the understanding of couple's gendered dynamics and how they may influence fertility desires. Such knowledge informs efforts to increase both spouses' involvement in reproductive decision-making and behaviors.