AI News Archive: June 30, 2026 — Part 18
Sourced from 500+ daily AI sources, scored by relevance.
- Beyond Binary Instrument QA: Probing Instrument Grounding in Music Audio-Language Models
Recent music audio-language models achieve high accuracy on instrument question-answering benchmarks, but it remains unclear whether this reflects robust audio grounding or benchmark-specific shortcuts. In this paper, we introduce an OpenMIC-derived diagnostic benchmark sequence for instrument groun...
- SwiftAudio: Data-Efficient Caption-Only Distillation for One-Step Text-to-Audio Diffusion-based Generation
Diffusion-based text-to-audio (TTA) models achieve impressive synthesis quality but suffer from high inference latency due to iterative multi-step denoising. Existing one-step approaches alleviate this issue but still rely on paired text--audio data during distillation. To address these limitations,...
- UniSAE: Unified Speech Attribute Editing on Speaker, Emotion and Low-Level Content via Discrete Phonetic Posteriorgram Modelling
Speech editing aims to modify specific portions of an utterance while preserving the remaining speech. Existing approaches primarily focus on word-level content modification and typically treat content, speaker, and emotion editing as separate tasks, limiting both editing granularity and flexibility...
- Reference-Based Prosody and Rhythm Evaluation for Spoken Dialogue Systems
Speech-to-speech (S2S) AI agents are advancing rapidly, yet evaluation lacks interpretable speech-native measures for conversational prosody and rhythm. Because $F_0$, speaking rate, articulation rate, and pausing shift with model-predicted speaker traits and interaction state, pooled human statisti...
- Usage frequency and application variety of research methods in library and information science: Continuous investigation from 1991 to 2021
The present study analyzed over 26,000 research articles published between 1991 and 2021 in twenty-one major LIS (Library and Information Science) journals, using the machine learning (ML) approach to categorize the research methods used by LIS scholars. The findings of this study are significant. F...
- Unsupervised Data-Efficient Cross-Modal Retrieval with Global-Neighborhood Alignment Hashing
Compared to supervised cross-modal hashing (CMH), unsupervised CMH reduces the reliance on manual labeling by learning binary codes from unlabeled image-text pairs. However, existing unsupervised CMH methods often rely on large-scale image-text pairs, which are costly to collect. To address this lim...
- One Retrieval to Cover Them All: Co-occurrence-Aware Knowledge Base Reorganization for Session-Level RAG
RAG systems retrieve documents optimized for answering one query at a time. Yet enterprise users arrive with sessions, that is, coherent episodes of related questions that span semantically distant parts of the knowledge base. We show that a single retrieval call over a standard knowledge base cover...
- GenPage: Towards End-to-End Generative Homepage Construction at Netflix
We present GenPage, an end-to-end generative approach to Netflix homepage construction that replaces the traditional multi-stage recommender stack with a single transformer. GenPage treats the user and request context as a prompt, and autoregressively generates the entire structured, multi-row homep...
- NLP Framework for Automated Symptom Severity Staging in Heart Failure and COPD Clinical Notes Using Ontology Integration: A Study Protocol
Background: Heart failure (HF) and chronic obstructive pulmonary disease (COPD) are among the leading causes of morbidity and mortality globally, with effective management heavily dependent on accurate severity staging using the New York Heart Association (NYHA) and Global Initiative for Chronic Obstructive Lung Disease (GOLD) classification systems. However, severity information is frequently embedded within unstructured clinical narratives rather than standardized Electronic Health Record (EHR) fields, limiting automated clinical decision support, disease surveillance, and retrospective healthcare analytics. Existing Natural Language Processing (NLP) approaches primarily rely on rule-based keyword extraction or supervised deep learning methods requiring large annotated corpora, which are often unavailable in many healthcare settings. Equally, most current systems inadequately integrate clinical ontologies for semantic reasoning and explainable classification, limiting interoperability and clinical applicability. Objective: This study aims to develop and evaluate an ontology-integrated NLP framework for automated extraction and severity staging of HF and COPD symptoms from de-identified clinical notes using NYHA and GOLD classification systems. Methods: The study will employ a Design Science Research (DSR) methodology to design, implement, and evaluate a hybrid NLP framework integrating rule-based extraction, SNOMED-CT ontology reasoning, and a Bidirectional Long Short-Term Memory with Conditional Random Field (Bi-LSTM-CRF) deep learning architecture for clinical sequence labeling. Approximately 1,000 de-identified clinical notes will be sampled proportionately from publicly available repositories including MIMIC-III/IV, eICU Collaborative Research Database, AmsterdamUMCdb, and MTSamples. Clinical text preprocessing will include tokenization, lemmatization, dependency parsing, abbreviation expansion, and negation detection. Ontology-guided semantic normalization will map extracted symptom entities to standardized SNOMED-CT concepts to support severity staging. Framework performance will be evaluated using precision, recall, F1-score, Cohens Kappa, sensitivity, specificity, positive predictive value, negative predictive value, confusion matrices, and correlation analyses against confirmed diagnoses and guideline-based severity classifications. Expected Outcomes: The proposed framework is expected to automate NYHA and GOLD severity staging across heterogeneous clinical note types without reliance on manually annotated severity labels. The ontology-integrated architecture is anticipated to improve semantic consistency, interpretability, and explainability of NLP outputs while enhancing EHR analytics, retrospective clinical audit, and AI-assisted clinical decision support. Conclusion: Findings from this study may provide a scalable and transferable framework for automated severity classification in data-rich but label-poor healthcare environments.
- Enhancing Title and Abstract Priority Screening Through SimEd AI Pipeline.
The exhaustive identification of evidence is central to systematic reviews, but the screening of titles and abstracts remains particularly labor intensive. Priority screening, an active learning approach that ranks records by estimated relevance, has emerged as an effective strategy to reduce screening workload. Its efficiency is commonly quantified using work saved over sampling at 100% recall (WSS@100%), representing the percentage reduction in effort compared with random screening. Although modern priority-screening models achieve high efficiency on many benchmark datasets, some reviews still exhibit low WSS@100%, indicating suboptimal retrieval. Our study sought to improve the retrieval of all relevant articles in challenging datasets to ensure better generalization of priority screening. We first showed using SYNERGY benchmark datasets that while the most advanced ELAS_h3 priority screening model from state-of-the-art ASReview LAB v.2 open-source software, efficiently retrieved most relevant articles, it struggled with the rare, final ones in challenging datasets. To address this, we tested a hybrid approach entitled SimEd AI: using ELAS_h3 for early retrieval and then applying supervised fine-tuning to the biomedical transformer BioMed-RoBERTa-base with these relevant articles to enhance the detection of the remaining difficult cases. We found that fine-tuning BioMed-RoBERTa-base model with 10 late-identified relevant and 10 hard irrelevant study titles and abstracts, enabled faster retrieval of articles of interest compared to ELAS_h3 alone. This approach increased WSS@100% from 46.5% (SD0.0%) to 83.3% (SD0.4%), while adding only an average of 22 minutes of computational time for fine-tuning and inference. The SimEd AI priority screening pipeline could be valuable for situations requiring highest possible recall. It could be particularly useful in scoping reviews with broad or diverse topics where traditional priority screening methods may miss subtle relevance signals. Further work should define a data-driven stopping rule for ending screening once the fine-tuned domain-specific transformer is applied at the final stage and assess generalizability across additional challenging datasets.
- Dynamic Graph Representation Learning for Data-Driven Huntington's Disease Staging: Evaluation Against Existing Embedding Methods and State-Space Models
Huntington's disease (HD) presents a heterogeneous neurodegenerative course, with motor, cognitive, and functional symptoms progressing differently across individuals. This atypical progression complicates the definition of discrete disease stages, hindering understanding of disease trajectories, timely pa- tient care, and therapy development. Consequently, current clinical staging systems rely heavily on clinician-defined, domain-specific criteria and fixed clinical measurement boundaries for stage assignment, reducing objectivity and often leading to overlapping clinical measurements across stages. While machine learning methods can help, existing approaches cannot fully capture complex temporal relationships within and across patients. We propose URL- STFN, a dynamic graph-based representation learning model that encodes both inter- and intra-patient temporal patterns from longitudinal clinical measures. We then evaluate disease stages formed through clustering and stability analysis of URL-STFN latent representations, and compare them with representations obtained from conventional embedding approaches. We further benchmark these clustering-based stages against states derived from conventional temporal models, including DHMM. We hypothesize that clustering URL-STFN latent representations enables identification of HD stages with reduced overlap in clinical measurements. The proposed framework is evaluated using 1,477 clinical visits from the Enroll-HD dataset, a large lon- gitudinal cohort with repeated clinical assessments. For staging, we used 44 clinical measurements spanning motor, cognitive, and functional domains. URL-STFN identifies clinically meaningful HD stages consistent with estab- lished disease progression while reducing overlap in clinical feature values compared with DHMM-derived and clinical staging approaches. These find- ings highlight the potential of a dynamic graph-based representation learning and clustering framework to support more objective, data-driven, and precise HD staging.
- Decoding the regulatory genetic architecture of endometriosis using AlphaGenome
Background Endometriosis is a complex, estrogen-dependent disease with a strong genetic component. Although genome-wide association studies (GWAS) have identified multiple susceptibility loci, most associated variants reside in noncoding regions, limiting biological interpretation and causal gene identification. Moreover, GWAS gene prioritization is limited by incomplete tissue-specific annotation coverage (e.g., GTEx, ENCODE, fine-mapping, Mendelian randomization, and network-based methods). We therefore applied the AlphaGenome artificial intelligence framework to prioritize endometriosis-associated variants based on predicted uterus-specific regulatory effects. Methods We analysed the top 10,000 endometriosis-associated single-nucleotide polymorphisms (SNPs) identified by previously published GWAS by Rahmioglu et al, using AlphaGenome across multiple genomic output types. Uterus-specific predictions with high-confidence effects (quantile score| [≥] 0.90) were grouped into major regulatory modalities. AlphaGenome-prioritized SNPs within {+/-}500 kb of known GWAS loci were classified into tiers based on the number of supported regulatory modalities, with broader support indicating stronger multilayer regulatory evidence. Effect allele frequency, linkage disequilibrium (LD), and overlap with previously published endometriosis-associated variants were also assessed. Results AlphaGenome generated uterus-specific, 147,033 high-confidence signals across 10,000 endometriosis-associated variants, spanning six regulatory modalities including gene expression, promoter activity, chromatin accessibility, transcription factor binding, histone modification, and RNA splicing. Within the 42 established endometriosis GWAS loci, AlphaGenome identified 42 alternative sub-threshold SNPs with stronger predicted uterus-specific regulatory effects than the published GWAS lead variants. Nineteen AlphaGenome-prioritized SNPs were classified as tier 1, showing support across all six regulatory modalities, compared with five GWAS lead SNPs. Linkage disequilibrium analysis identified eight tier 1 SNPs with weak-to-low LD (r<2> < 0.5) relative to the corresponding GWAS lead variants, regulating majority of genes involved in estrogen-driven proliferation and inflammatory signalling, highlighting their potential relevance to endometriosis pathogenesis. Additionally, we identified 167 genome-wide significant SNPs outside 42 published GWAS lead SNP loci including six tier 1 SNPs (rs1482061, rs7772579, rs6557140, rs2982571, rs12631337 and rs79626929), encompassing genes nearby ESR1/6q25.1, substantiating biological relevance for endometriosis pathogenesis. Conclusions AlphaGenome-based regulatory prioritization refined endometriosis-associated genome-wide association study loci by identifying variants with stronger predicted uterus-specific functional relevance. These findings provide a regulatory framework for prioritizing candidate variants and genes for downstream functional validation in endometriosis.
- Infoxmed2.0-27B: Instruction Tuning, Preference Alignment, and GRPO-Based Reward Model Training for Medical LLMs
Abstract-Large language models (LLMs) [1], [2] have demon strated remarkable capabilities across general domains, yet their application in specialized medical contexts demands rigorous domain adaptation [3], [4]. We present Infoxmed2.0-27B, a medical foundation model built upon Qwen3.5-27B [5] through a comprehensive multi-stage post-training pipeline: (1) proprietary medical data synthesis from a MySQL database with MedicalCategoryTree organization, medical PhD team validation, Chinese RoBERTa [6] semantic deduplication, and API-assisted language refinement; (2) instruction supervised fine-tuning of Qwen3.5- 27B via LoRA [7] (r = 8, = 32) using MS-Swift [8], producing iterations Infoxmed2.0.0[->]2.0.2[->]2.0.4; (3) Direct Preference Optimization (DPO) [9] on 6,283 curated medical preference pairs [10] using DPO-RPO loss ({beta} = 0.3, RPO = 0.1) across eight progressive training iterations (v0-v7); and (4) parallel Group Relative Policy Optimization (GRPO) [11]-based medical reward model training on Qwen3.5 combining internal rule-based reward functions with external DeepSeek signals. Comprehensive evaluations under a uniform LLM-as-Judge [12] framework with GPT-5.4 demonstrate 77.0% accuracy (mean quality score +7.18) on MedMCQA [10] and +2.59 on HLE, with pipeline progression from +6.69 (base) to +7.06 (SFT) to +7.18 (final).
- Scalp EEG reveals functional dissociable aperiodic timescales in divergence of mental health
Aperiodic neural activity is increasingly used as a marker of brain health, yet it remains unclear whether this signal reflects a single neural timescale or separable slow and fast processes with distinct relevance to ageing and disease. Here, using resting-state scalp EEG from approximately 1,700 participants across healthy, neurological and psychiatric disorders and chronic-pain cohorts, we show that spectral knees recovered from individual spectra converge into two reproducible population-level components. These slow and fast aperiodic timescales showed distinct functional profiles: the slow component remained comparatively stable across healthy ageing, brain-related disorders and chronic pain, whereas the fast component increased with healthy ageing, decreased in brain-related disorders and was unchanged in chronic pain. These findings establish that scalp EEG preserves functionally dissociable slow and fast aperiodic timescales, possibly reflecting body-cognitive interactions, and suggest that the fast component may serve as a scalable, non-invasive marker of brain ageing and dysfunction.
- Health conditions and RSV-related Pediatric Intensive Care Unit admissions in children during their second RSV season
Importance: Respiratory syncytial virus (RSV) hospitalization rates are highest among children <2 years of age. RSV immunization with infant monoclonal antibody or maternal vaccine is recommended to protect all U.S. infants in their first RSV season. For certain high-risk children aged 8-19 months entering their second RSV season, the monoclonal antibody nirsevimab is recommended. Little is known regarding preexisting health conditions as risk factors for RSV-associated respiratory failure in children during their second season. Objectives: To describe children admitted to the pediatric intensive care unit (PICU) for RSV during their second RSV season by preexisting health conditions, and to compare demographic and clinical characteristics across groups. Design, Setting, and Participants: Surveillance registry of children 8- <24 months old admitted to the PICU in 30 pediatric hospitals in the 2023-2024/2024-2025 RSV seasons. All children had an RSV-positive respiratory sample and received respiratory support with high flow nasal cannula, noninvasive ventilation, or invasive mechanical ventilation (IMV). Exposure: Preexisting health conditions potentially increasing risk of severe RSV disease. Main Outcomes and Measures: Patients were classified into four mutually exclusive groups by preexisting health conditions: 1) U.S. nirsevimab eligible criteria, 2) other identified RSV risk conditions (with some evidence of increased risk for severe RSV), 3) other preexisting conditions, and 4) no preexisting conditions. Patient demographic characteristics and level of respiratory support received were compared. Results: Among 574 children: 47 (8.2%) had U.S. nirsevimab eligibility criteria, 76 (13.2%) had other RSV risk conditions, 96 (16.7%) had other preexisting conditions, and 355 (61.8%) had none. A higher proportion of children with nirsevimab eligibility factors (40.4%) than those with other identified RSV risk conditions (17.1%) required IMV, which was higher than other (10.4%) or no (5.9%) preexisting health conditions (ptrend<0.001). Conclusions and Relevance: Approximately 20% of children admitted to the PICU with severe RSV were in the defined groups that met U.S. nirsevimab-eligibility criteria or that had an identified RSV risk condition associated with known risk for severe RSV. A considerable proportion of both groups of children required IMV for respiratory support. These findings may help inform future deliberations regarding U.S. second season nirsevimab-eligibility recommendations.
- Transdermal Clonidine versus Spironolactone in Resistant Hypertension
Objectives: To compare real-world cardiovascular outcomes and safety events in patients with resistant hypertension following initiation of transdermal clonidine (TC) or spironolactone. Methods: A retrospective analysis was performed using Merative MarketScan(R) Databases in the USA to identify cohorts with resistant hypertension initiating TC or spironolactone as a fourth-line agent between January 2012 and September 2024. Major Adverse Cardiovascular Events (MACE) and safety events were assessed during variable follow-up periods. Inverse probability of treatment weighting (IPTW) was applied to adjust for differences in baseline characteristics. Cox proportional hazard models were used to adjust for post-index beta-blocker utilization as a time-varying covariate for MACE outcomes. Results: The analysis included 3,113 patients in the TC cohort and 30,640 in the spironolactone cohort. After IPTW, baseline characteristics were well balanced between cohorts (standardized mean differences <0.10; mean age 60 years, 54% male). Mean follow-up was 7.1 and 10.5 months for the TC and spironolactone cohorts, respectively. After IPTW no differences in MACE outcomes were observed between the two cohorts (weighted rate ratio 1.27 [0.79-2.06]). Results were consistent after adjusting for post-index beta-blocker use. The risk of hyperkalemia was significantly lower in the TC cohort (weighted rate ratio, 0.48 [0.33-0.70]. Conclusions: In this real-world analysis, patients with resistant hypertension treated with TC have similar risk for MACE outcomes as with spironolactone, but with significantly lower risk of hyperkalemia. Thus, in patients with resistant hypertension TC appears to provide similar cardiovascular protection, with a more favorable safety profile.
- FROM PRIORITIZATION TO ACCESS: DECISION PATHWAYS FOR ESSENTIAL HEALTH TECHNOLOGIES IN SOUTH AFRICA'S PUBLIC SECTOR
Background Timely access to essential health technologies depends on aligning evidence-informed adoption with financing, procurement, delivery, monitoring, and learning. We examined how decision pathways shaped selected health technologies progression from prioritisation to implementation in South African public sector, and their implications for access, equity, sustainability, and learning. Methods We conducted a qualitative, retrospective, multi-case health policy analysis of twenty essential health technologies purposively selected to capture variation in technology type, disease area, delivery platform, adoption trajectory, and implementation outcome. Document review, process mapping, and key informant interviews were used to reconstruct decision pathways. Analysis was guided by the Policy Cycle Framework, Health Policy Triangle, and Health Technology Assessment (HTA) process domains. Results Decision pathways varied by technology and delivery platform but followed a common sequence from prioritisation and appraisal to policy endorsement, implementation, and limited reassessment. Medicines and vaccines were generally embedded within established national decision structures. Diagnostics required coordination across laboratory, programme, procurement, and service-delivery systems, while medical-device decisions were more decentralised. Upstream appraisal focused on safety, effectiveness, and public health needs; affordability, infrastructure, equity and sustainability were addressed inconsistently. System learning was evident when routine data, pharmacovigilance, programme review, and guideline revision informed post-adoption adaptation. Weak feedback loops limited reassessment of implementation barriers, equity effects, and sustainability, contributing to delays despite policy endorsement. Conclusion South Africa has formal structures for evidence-informed technology adoption. HTA would be strengthened by treating appraisal as part of lifecycle governance, with earlier alignment between adoption decisions, financing, procurement, implementation readiness, monitoring, and reassessment.
- Shared trans-ancestry architecture of HLA-mediated disease risk in the All of Us Research Program
The human leukocyte antigen (HLA) region is the strongest genetic contributor to many immune-mediated diseases, yet whether HLA architecture is shared across ancestries remains unclear. We analyzed high-resolution HLA variation in 390,823 participants from the All of Us Research Program spanning six genetic ancestry groups, including 262,915 with linked electronic health records. Using whole-genome sequencing and graph-based inference, we genotyped 20 HLA genes at G-group resolution and identified 4,780 distinct alleles. Analyses accounting for disparate sample sizes demonstrated that ancestry-private allelic variation reflected unequal discovery depth rather than ancestry-population specificity. A meta-analysis of ancestry-stratified phenome-wide association analyses with 363 HLA alleles with frequency > 0.001 and 3,430 clinical phenotypes identified 1,461 significant HLA-phenotype associations (FDR < 0.05). Although many associations reached significance in only one ancestry group, effect directions were largely concordant, highlighting differences in allele frequency, linkage disequilibrium, and statistical power among ancestry groups. Stepwise conditional modeling demonstrated that common complex trait variation could be concurrently explained by five to seven independent HLA allele signals. These findings demonstrate that a multi-ancestry, phenome-wide study can distinguish true biological heterogeneity from sampling-driven detectability differences in HLA.
- Genetic and transcriptomic determinants of disseminated coccidioidomycosis risk reveal African ancestry enrichment and rare immune response variants in NLRX1
Coccidioidomycosis, also known as Valley Fever, is a fungal disease endemic to the Americas that kills hundreds annually, yet the host factors that lead to increased risk of life threatening dissemination of coccidioidomycosis, remain poorly understood. We assembled the largest, comprehensively sequenced coccidioidomycosis cohort to date, comprising 795 individuals with laboratory confirmed coccidioidomycosis and clinical disease severity phenotyping, many with paired whole blood genomic and transcriptomic data. Individuals with greater than 50% African genetic ancestry are significantly enriched in disseminated coccidioidomycosis (DCM) cases (OR=13.37, p=1.08 x10-18), reflecting ancestry-associated differences in allele frequencies at immune loci. Transcriptomic profiling (n=267) revealed upregulation of interferon-inducible genes IFI44 and IFI44L, the fungal recognition receptor CLEC4D, and pro-inflammatory protein S100A12, with sex-specific expression differences in immune cell composition. Gene-burden testing identified NOD-like receptor NLRX1 as the only gene carrying significantly more damaging rare variants than expected by chance (p=5.85 x10-4). We identified a rare missense variant, p.Arg252Trp (rs145644388), in five patients with DCM that represents a founder variant: all carriers share African genetic ancestry and carry 0.6-1.1 centimorgans of identical-by-descent sequence, indicating origin from a common ancestor. In gnomAD, p.Arg252Trp shows a 47-fold enrichment in African populations compared to European populations, directly linking this rare variant to the population-level African ancestry enrichment in DCM. We also identified a rare variant in IFI27 (p.Ser63Leu) present exclusively in severe coccidioidomycosis (DCM or CPC), with all five carriers being male. NLRX1 disruption impairs LC3 associated phagocytosis, a key antifungal mechanism in macrophages. Together, these findings reveal both ancestry-linked immune dysregulation and rare-variant architectures, including a novel African ancestry-enriched founder variant, underlying severe coccidioidomycosis, and identify new targets for risk stratification and treatment.
- scEPS integrates genetic and single-cell disease atlas data to provide granular mechanistic insights into complex human diseases
Integrating GWAS and single-cell data holds great potential for prioritizing causal disease biology at cellular resolution. Recent integrative approaches typically assess the enrichment of disease genetic signals in cell types or individual cells, without directly modeling disease phenotypes. We develop a new method, single-cell Expression exPlainability Statistics (scEPS), for identifying disease-associated cell neighborhoods, by explicitly testing whether the expression of GWAS-prioritized genes explains more variance in a disease than randomly selected, mean-expression-matched control genes. Crucially, when applied to PRSs of healthy donors, scEPS captures the genetic covariance between gene expression and diseases, mitigating the effect of reverse causation and prioritizing cell populations mediating the effects of GWAS genes. We applied scEPS to clinical diagnoses and PRSs of 4 neurological and 4 respiratory disorders, integrating brain and lung cell atlas data, respectively, with respective GWAS summary statistics data. scEPS recapitulated known and uncovered novel disease-associated cell populations, identifying 1.77x (s.e. 1.21) and 5.13x (s.e. 3.08) more significant associations than a CNA-based approach and scDRS, respectively. Furthermore, scEPS detected different cell populations, contrasting clinical diagnoses vs. their PRSs, revealing distinct biology for the active/symptomatic vs. preclinical/asymptomatic states of the disease. Finally, we observed limited concordance across methods using distinct definitions of disease association, underscoring the need to integrate complementary insights for holistic understanding of disease biology.
- Evaluating Polygenic Score Transferability for Lipid Traits in Underrepresented Populations: Evidence from Samoan Cohorts
Dyslipidemia is a significant risk factor for cardiovascular disease (CVD), the leading cause of death in Samoa, accounting for 34% of deaths. Polygenic scores (PGS) derived from large scale multi ancestry genome-wide association studies offer potential for improved CVD risk prediction by aggregating genetic effects on lipid traits, yet their performance in Pacific Islander populations remains largely unknown. We evaluated the transferability of multi-ancestry PGS for LDL cholesterol (LDL C), HDL cholesterol (HDL C), triglycerides (TG), and total cholesterol (TC) in 4,342 Samoan adults across five cohorts spanning 1990 to 2010. PGS derived from Graham et al. and Kanoni et al. multi-ancestry meta-analyses were harmonized with genome-wide imputed genotypes using a Samoan-specific reference panel, and performance was assessed using incremental R^2 from linear mixed models with bootstrapped confidence intervals. PGS performance varied across traits and cohorts: HDL C showed the highest performance (incremental R^2 5.0 to15.0%), followed by LDL C (5.7 to 8.6%) and TC (5.0 to10.7%), with TG showing the lowest performance (3.5 to 7.0%). Meaningful LDL C transferability was achieved only when using a genome-wide PRS CS score (99.6 to 99.7% variant matching), whereas a curated pruning-and-thresholding score achieved only ~9% matching and near-zero performance. These findings establish the first systematic benchmarks for lipid PGS performance in Samoans, demonstrate that multi-ancestry scores can achieve meaningful transferability in this underrepresented population when genome-wide variant coverage is ensured, and highlight the importance of rigorous variant harmonization assessment prior to clinical deployment of PGS in diverse populations.
- Real-world activity of trastuzumab deruxtecan in heavily pretreated HER2-expressing ovarian cancer: focusing on HER2-low responses and CCNE1 amplification
Background: Trastuzumab deruxtecan (T-DXd) is active in HER2-expressing solid tumours, but trials excluded HER2 immunohistochemistry (IHC) 1+ disease, and data in pretreated ovarian cancer are lacking. We evaluated real-world T-DXd activity and genomic correlates in pretreated ovarian cancer, predominantly high-grade serous (HGSOC). Methods: HER2 expression was assessed in an unselected ovarian cancer cohort (N=74). Fifteen patients receiving off-label T-DXd (14 HGSOC, 1 clear cell; IHC 1+ to 3+) had HER2 status centrally confirmed using gastric-type criteria. Activity was assessed by intra-patient growth modulation index (GMI; progression-free survival [PFS] on T-DXd divided by PFS on the prior line; [≥] 1.33 considered meaningful). Patients on treatment at data cut-off were censored. Objective response (RECIST 1.1) was assessed centrally where imaging was available (n=8). Results: Of the 40 HER2-expressing tumours, 15 received T-DXd, limited mainly by reimbursement. Among 14 evaluable patients (median 5 prior lines), 9 reached a GMI [≥] 1.33 (median 1.69); 8 remained on treatment at cut-off, making durability preliminary. Confirmed partial responses occurred across the HER2 spectrum. Benefit was independent of homologous-recombination (HR) status: one HR-proficient, CCNE1-wild-type patient achieved prolonged control and was rendered disease-free after radiotherapy to an oligoprogressive lesion. Exploratory analysis showed all four evaluable CCNE1-amplified tumours had reduced or non-durable benefit. Conclusions: T-DXd shows preliminary, clinically meaningful activity in HER2 IHC 1+ ovarian cancer independent of HR status. CCNE1 amplification may attenuate benefit, a candidate biomarker for WEE1-inhibitor combinations. Approval restricted to IHC 3+ disease would exclude most responders in this cohort. Prospective validation is required.
- Biomechanical Analysis of Dynamic Gripping in Manual Laborers Exhibiting Work-Related Scapholunate Instability Signs
ABSTRACT Background: Scapholunate instability (SLI) represents the most prevalent form of carpal instability and is increasingly recognized as a clinically significant occupational condition in manual-labor populations. Repetitive forceful gripping, sustained wrist loading, and awkward joint postures inherent to construction, manufacturing, and heavy industry may predispose workers to progressive scapholunate ligament compromise; however, the specific biomechanical mechanisms underlying dynamic grip force transmission in symptomatic laborers remain poorly characterized. Objective: This study aimed to quantify dynamic grip force profiles and wrist kinematics in manual laborers exhibiting clinical signs of scapholunate instability; compare these biomechanical parameters with asymptomatic worker controls; and identify occupation-specific loading patterns that may contribute to scapholunate ligament compromise. Methods: A cross-sectional observational design was employed. Forty-two male manual laborers, mean age 36.4 (SD 7.2 )years; mean occupational exposure 9.8 (SD 4.1) years, were recruited from construction and surgical instrument manufacturing sites in Sialkot District, Pakistan. Participants were stratified into a symptomatic group (n = 21) based on positive Watson scaphoid shift test, dorsal wrist pain, and functional limitation, and an asymptomatic control group (n = 21). Dynamic grip force was measured using a calibrated Jamar dynamometer across five standardized trials. Wrist kinematics were captured during a simulated gripping task using electrogoniometry. Surface electromyography (sEMG) recorded flexor digitorum superficialis, flexor carpi ulnaris, and extensor carpi radialis longus activity. Scapholunate gap was confirmed radiographically. Statistical analyses included independent samples t-tests, Pearson correlation, and logistic regression ( = 0.05). Results: Symptomatic workers demonstrated significantly reduced grip force (28.4 (SD 5.6) kg versus 41.7 (SD 6.3) kg; p < 0.001) and elevated wrist ulnar deviation during peak gripping (22.3 (SD 4.1) degrees versus 14.6 degrees (SD 3.8) degrees ; p = 0.002) compared with controls. sEMG amplitude of flexor digitorum superficialis was paradoxically elevated in the symptomatic group (mean difference 18.3 microvolts; 95% CI 11.4-25.2), suggesting compensatory hyperactivation. Scapholunate gap correlated positively with years of occupational exposure (r = 0.67; p < 0.001). Logistic regression identified wrist ulnar deviation angle (OR = 2.14; 95% CI 1.43-3.19) and grip force asymmetry (OR = 1.87; 95% CI 1.21-2.89) as independent predictors of instability signs. Conclusions: Manual laborers with work-related scapholunate instability signs exhibit distinct biomechanical signatures during dynamic gripping, including reduced force output, abnormal wrist posture, and compensatory neuromuscular recruitment. Occupational exposure duration is a significant predictor of radiographic scapholunate dissociation. These findings support the integration of ergonomic risk screening and early biomechanical assessment into occupational health protocols for manual labor populations in low- and middle-income settings. Abbreviations: CI: confidence interval, DRUJ: distal radioulnar joint, EMG/sEMG: (surface) electromyography, FCU: flexor carpi ulnaris, FDS: flexor digitorum superficialis, MSD: musculoskeletal disorder, OR: odds ratio, ROM: range of motion, SL/SLI: scapholunate/scapholunate instability, SLIL: scapholunate interosseous ligament, SLAC: scapholunate advanced collapse, WMSDs: work-related musculoskeletal disorders, SD: standard deviation. Keywords: scapholunate instability; dynamic gripping; manual labor; wrist biomechanics; carpal kinematics; occupational ergonomics; grip force; electromyography; Watson scaphoid shift test; work-related musculoskeletal disorders
- Nucleus-specific thalamic involvement in seizure networks differentiates neuromodulation outcomes
Closed-loop neuromodulation via responsive neurostimulation (RNS) of the thalamus has emerged as a promising therapy for drug-resistant epilepsy (DRE), particularly in patients with broad or multifocal onset. However, response to thalamic RNS is inconsistent, and there is a crucial need to identify factors that distinguish responders from non-responders. Given the heterogeneous composition of the thalamus, the specific contributions of individual thalamic nuclei during seizures may explain the variability in outcomes between patients and could potentially serve as biomarkers for guiding target selection. We analyzed 129 seizures from 28 patients with DRE who underwent stereo-EEG monitoring with recordings of the centromedian (CM: n = 15) or pulvinar (PLV: n = 13) thalamic nuclei and were subsequently treated with RNS targeting the corresponding nucleus (CM: 11/15 [73%] responders; PLV: 7/13 [54%] responders). Patients were classified as responders (Engel class I-III) or non-responders (Engel class IV) based on reduction in seizure frequency. For each seizure, we constructed functional connectivity networks spanning seizure onset to termination and quantified the role of the thalamic nucleus by computing its total node strength. We also used an automated detection algorithm to measure the time of seizure spread to each thalamic nucleus relative to seizure onset. Connectivity and spread timing were then compared between responders and non-responders within each nucleus group. The timing of thalamic recruitment following seizure onset did not differ significantly between responders and non-responders in either nucleus, although CM responders showed a non-significant trend toward earlier recruitment. Analysis of functional connectivity revealed nucleus-specific patterns. CM responders exhibited significantly higher thalamic node strength than non-responders during the late-seizure phase, with no significant difference at early- or middle-seizure phases. PLV responders showed significantly higher thalamic node strength during the middle-seizure phase, but there was no significant difference at early- or late-seizure phases. These findings suggest that the degree and timing of thalamic involvement during seizures may serve as biomarkers for predicting response to thalamic RNS in DRE. CM involvement in responders was characterized by stronger connectivity that persisted through seizure termination, whereas PLV involvement in responders was reflected primarily in connectivity during seizure propagation and progression. Incorporating these nucleus-specific ictal network features into pre-surgical evaluation could improve patient selection and guide nucleus-specific targeting for thalamic RNS.
- Early immune activation in the prediagnostic phases of immune-mediated neurological diseases
Multiple sclerosis (MS), myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD), and neuromyelitis optica spectrum disorder (NMOSD) are immune-mediated inflammatory disorders of the central nervous system (CNS). The temporal relationship between disease-specific autoantibodies and biomarkers of CNS injury before diagnosis remains unclear and is relevant for understanding early pathobiology. Here, we conducted a multicentre retrospective longitudinal case-control study using prediagnostic plasma from 362 individuals who later developed MS, 145 who developed MOGAD, and 60 who developed NMOSD. Plasma IgG levels against CNS antigens, MOG, and AQP4, as well as neurofilament light chain (pNfL), were quantified, and temporal relationships between immune activation, neuroaxonal injury, and clinical disease onset were modelled using linear mixed-effects models and survival analyses. In MS, EBNA-1-specific and CNS-cross-reactive IgG were elevated up to 77.8 months before diagnosis, preceding pNfL increases by 44.9 months. In NMOSD, AQP4-IgG seroconversion occurred 32.5 months before diagnosis and preceded pNfL elevations by 40.4 months. In MOGAD, pNfL elevations preceded MOG-IgG seroconversion by 11.2 months. Thus, in MS and NMOSD, humoral autoimmunity precedes detectable CNS injury, whereas in MOGAD, neuroaxonal injury occurs before circulating MOG-IgG. These distinct temporal patterns suggest differing early immunopathological trajectories and may provide a framework for future studies of early disease biology and biomarker-guided risk stratification.
- CSF proteome-wide study of neuropsychiatric symptoms of dementia
Introduction: Neuropsychiatric symptoms in dementia (NPS) are common and among the most troubling aspects of living with dementia, yet their underlying mechanisms remain unclear. Here, we aimed to identify cerebrospinal fluid (CSF) proteins associated with NPS. Methods: Proteomes were profiled from CSF collected at baseline from participants of the Alzheimer's Disease Neuroimaging Initiative (ADNI) using mass spectrometry. Here, we included participants having positive AD CSF biomarkers (i.e., pTtau181 / Abeta42 ratio >0.025) and mild cognitive impairment or AD dementia (n=419). Eight NPS domains were assessed longitudinally with the Neuropsychiatric Symptom Inventory Questionnaire. Severity of cognitive impairment was evaluated using the CDR-SB. Proteome-wide differential expression analysis for each NPS domain at baseline was performed. Significant protein-NPS associations underwent mediation analysis to test whether they were mediated by cognitive impairment severity. Cox proportional hazard was modeled for baseline CSF proteins and incident NPS. Additionally, we tested whether candidate NPS causal proteins previously identified in brain are associated with NPS in CSF. Results: We identified 8 CSF proteins associated with apathy at baseline (FDR q<0.05) after adjusting for sex, age, and education - NTNG2, S100A1, FZD1, FSTL5, CDH7, CHODL, FBXO2, and CACNA2D2. Mediation analysis revealed these associations were independent of cognitive impairment severity in four proteins and only partially mediated by cognitive impairment severity in the remaining four proteins. Among the 10 NPS candidate causal proteins previously identified in brain and detected in CSF, the abundance of two proteins (CPD, GRN) in CSF was associated with baseline disinhibition and of two other proteins (PIK3IP1, PCMT1) with both baseline apathy and incident apathy after adjusting for sex, age, and education. Discussion: These findings suggest that proteomic alterations in apathy in MCI/AD encompass synaptic connectivity, calcium regulation, Wnt-signaling, and neuronal proteostasis, highlighting potential CSF biological processes and biomarker candidates for apathy.
- NSAID use is associated with lower dementia and Alzheimer disease prevalence and slower cognitive decline: A retrospective longitudinal analysis of the NACC cohort
INTRODUCTION: Dementia, particularly Alzheimer disease (AD), is a major global health challenge, with prevalence projected to reach 150 million cases by 2050. AD is characterized by progressive cognitive decline linked to neuroinflammation and neurodegeneration. Non-steroidal anti-inflammatory drugs (NSAIDs) have been explored as potential neuroprotective agents, particularly diclofenac, which has been proposed to modulate microglial inflammasome signaling. However, prior studies investigating NSAIDs in AD have yielded inconsistent findings. We therefore reexamined the relationship between selected NSAIDs and dementia outcomes in a large longitudinal cohort from the National Alzheimer Coordinating Center (NACC). METHODS: We analyzed cross-sectional and longitudinal data from the NACC database collected between 2005 and 2022. Associations between NSAID exposure and dementia, AD, and cognitive trajectories were examined. Propensity score matching was performed to compare NSAID users with matched non-users while adjusting for demographic and clinical confounders. Longitudinal mixed-effects models were used to assess cognitive decline based on Montreal Cognitive Assessment (MoCA) scores. RESULTS: Among 47,165 participants, diclofenac and naproxen use were associated with a lower prevalence of dementia and AD compared with matched non-users, whereas etodolac showed no significant associations. Diclofenac users demonstrated reduced odds of dementia and AD. Naproxen showed similar cross-sectional associations. In longitudinal modeling, diclofenac users had a significantly slower rate of cognitive decline than non-users. DISCUSSION: These findings suggest a compound-specific association between NSAID use and AD, with diclofenac potentially modulating disease progression through anti-inflammatory mechanisms. The observed modulation of longitudinal cognitive decline supports further investigation of inflammatory pathways, including microglial and inflammasome signaling, as therapeutic targets in biomarker-defined AD populations.
- Using MRI whole brain atrophy and clinically reported outcomes in combination to assess interim treatment response in multi-arm multi-stage trials in progressive multiple sclerosis
Background: Interim stage outcomes in multi-arm multi-stage trials need not be the same as the final primary outcome, and should be selected with the goal of providing the best chance of continuing with an effective treatment in the study during the early stage where there is also potential to drop an ineffective arm. Jointly considering multiple outcomes can enhance that ability to detect an emerging signal. Methods: The Optimal Clinical Trials Platform for Progressive Multiple Sclerosis (OCTOPUS) study is a randomised, placebo-controlled, double-blind, phase 3, MAMS trial testing treatments for people with progressive multiple sclerosis. The interim analysis was to be based solely on an MRI outcome; reduction in whole brain atrophy rate. Accumulating data from external trials led to concern that this outcome may result in prematurely rejecting an effective treatment. As a solution we propose adding 3 clinical outcomes to the MRI outcome and propose a multivariate mixed model to accommodate them jointly, despite significant differences in scale and even measurement type. We show how use of the model-derived covariances allows us to linearly combine the treatment effects of disparate outcomes into a single treatment effect. We also describe how to analytically calculate power for this combination and compared it to the individual components' performance. Results: Based on variance data from a previous study, we found power was increased moderately by 7%, from 83% to our target 90% given the assumed respective treatment effect sizes for OCTOPUS. When considering observed effect sizes from other trials these power improvements were maintained, despite there being great variability between the outcome effects. Conclusions: Whilst not greatly boosting power, we argue that this strategy improves the interim outcome measure by also making it more resilient to the uncertainty surrounding effect size, and mitigating against unexpected negative results by spreading the liability across related but distinct outcomes.
- PCA-Guided Separation of Mixed Motor Unit Sources in High-Density EMG
Objective: Decomposition of high-density electromyographic signals enables non-invasive analysis of individual motor unit (MU) behavior, but reliable interpretation of physiological changes in health and disease depends on accurate MU discharge detection. This accuracy is compromised by mixed source estimates, where high amplitude peaks are associated with discharges from more than one MU. We introduce a post-decomposition framework to identify and separate suspected mixed sources using PCA-guided source refinement. Method: For each suspected mixed source, extended and whitened EMG vectors were extracted at source peaks and projected into a low-dimensional PCA subspace. This subspace highlighted MU-specific differences across candidate discharges, including subtle or spatially localized features of the spatiotemporal MUAP profile. Clusters in the PCA subspace were used to initialize source estimates for the constituent MUs. During iterative source refinement, source peak amplitudes were reweighted according to the distance of their corresponding points from the associated cluster center. Particle swarm optimization selected the reweighting factor that minimized the coefficient of variation of inter-spike intervals (CoVISI). Results: The algorithm separated mixed MU sources in simulated and experimental HDsEMG data. In simulated data, resolving mixed sources increased median rate of agreement (RoA) by >40%. In experimental recordings, MU yield increased by 1.27 per trial and CoVISI decreased by 0.28 (33% RoA improvement). Conclusions: PCA-based representation enhanced separability between MUs with similar MUAP profiles, while distance-based amplitude reweighting reduced re-merging during source refinement. Significance: This framework resolves merged MU discharge trains, improving decomposition accuracy and recovering MUs that might otherwise be excluded by quality thresholds.
- Comprehensive Demographic Correction Improves Sensitivity and Reduces Bias in Cognitive Assessment
Background. Scores on neuropsychological assessments are typically corrected for the influences of age, education, and gender (AEG). However, other demographic factors, such as crystallized ability and race/ethnicity, independently affect test performance. As a result, standard scores systematically over- or under-classify impairment in patients whose demographic profile differs from that of the reference population. Methods. We developed a Comprehensive (C-) model scoring algorithm that added vocabulary, age-squared, race/ethnicity, Latino background, a coarse socioeconomic status proxy, computer use, and daily prescription medications to the standard AEG predictor pool. The model was developed using data from 1,914 community-dwelling adults assessed with the California Cognitive Assessment Battery (CCAB; Woods et al., 2024). For each of 118 individual cognitive measures, stability-selection LASSO identified robust predictors in 300 random 80/20 splits retained at >=80% frequency and then estimated mean coefficients and confidence intervals in 1,000 bootstrap OLS samples. Cross-sample frozen-coefficient validation was used to evaluate scoring model generalization in two subgroups: Group 1 (n = 1,033, older, first enrolled cohort) and Group 2 (n = 881, a recently recruited younger cohort). Results. Stability selection retained a mean of 2.81 predictors per measure (range 1-6). Compared to the AEG model, the C-model approximately doubled variance explained (r2 = 0.50 vs 0.25; mean across cognitive domains r2 = 0.32 vs 0.18) and outperformed AEG in 98.8% of individual measures with non-trivial demographic signal. Racial disparities in MCI classification (the bottom-7th-percentile) were substantially reduced: Black-vs-White ratios fell from 5.6 (AEG) to 1.8 (C). Conversely, sensitivity was improved in individuals with elevated premorbid function: MCI classification ratios in low-vs-high vocabulary quartiles fell from 11.3 to 2.1. AIC favored the C-model in 88.1% of measures (mean delta-AIC = -167), ruling out overfitting. Frozen-coefficient validation preserved the C-model's r2 advantage in every cognitive domain. Conclusions. By correcting scores for race, premorbid cognitive functioning (vocabulary), and other demographic predictors, the C-model explains substantially more variance than the AEG model, reduces racial bias, and increases sensitivity to cognitive decline in high-functioning participants. C and AEG models can be used in parallel: model concordance increases diagnostic confidence, while disagreement carries diagnostic information.
- Prevalence of Parkinson's disease in Lagos, Southwestern Nigeria: a descriptive community-based study from the Transforming Parkinson's Care in Africa (TraPCAf) project.
Background The global burden of Parkinson's disease (PD) has increased substantially over recent decades, driven by population ageing and rising age-standardized prevalence. In Africa, accurate estimates remain limited due to a lack of recent, methodologically robust population-based studies. Objectives To determine the current age-standardized and sex-specific prevalence rates of PD in Nigeria. Methods We conducted a 2-stage, cross-sectional population-based door-to-door survey among adults aged [≥]18 years in two densely populated urban local government areas in Lagos State, Nigeria, between April 1, 2024 and January 31, 2025. The first stage involved a household census and screening for parkinsonism using a standardized screening tool. The second stage consisted of in-person clinical assessment and diagnostic confirmation by physicians using established clinical diagnostic criteria. Crude and age-standardized prevalence rates (to the World Health Organization World Standard and European Standard Populations) were calculated. Results 31,009 individuals (52.7% female) from 13,222 households were surveyed, and 70 persons were diagnosed with PD. The crude prevalence ratio was 225.7 per 100,000, with higher prevalence in males (53/14658, 361.6) than females (17/16,351, 104.0). The age-standardized prevalence rate (95% confidence interval) was 193 per 100,000 (150 -- 245) (females: 86 (50 -- 137); males: 277 (207 -- 362)), and increased with advancing age. The diagnostic gap (previously undiagnosed) was 60.0% (42/70). Treatment gap (never treated) was 44/70 (62.9%). Conclusions The age-standardized prevalence of PD is higher than previously reported in sub-Saharan Africa. These findings provide contemporary data to inform updated estimates of disease burden and support health systems planning.
- Prognostic value of plasma brain-derived pTau
Background: Plasma brain-derived pTau217(BD-pTau217) may provide a Alzheimers disease-specific plasma tau measure than total pTau217, but its prognostic value is unclear. We compared BD-pTau217 and total plasma pTau217 for predicting clinical and amyloid PET progression in cognitively unimpaired (CU) ADNI participants. Methods: Plasma NULISAseq biomarkers were measured in 1,427 ADNI participants, including 529 CU individuals. Amyloid PET progression was assessed in baseline CU amyloid-negative participants (Centiloid [≥] 24.1) with longitudinal PET imaging; clinical progression was assessed in all baseline CU participants. Associations were evaluated using Cox models and time-dependent AUC. Results: BD-pTau217 did not clearly outperform total pTau217 for predicting progression to mild cognitive impairment or dementia. However, among baseline amyloid-negative participants (N=175), BD-pTau217 better predicted amyloid PET positivity at 2.5 years (tdAUC 0.82 vs 0.69; HR=10.54, p=0.00015) and 4 years (tdAUC 0.77 vs 0.64; HR=7.03, p=0.00055). Conclusion: BD-pTau217 improved prediction of near-term amyloid PET progression, with less clear advantage for clinical progression.
- StemSplitterAI
Split any song into stems and custom backing tracks online.
- Epidemiology of diabetic foot in Bukavu
Abstract The prevalence of diabetes is increasing in all countries, and diabetic foot is a complication with serious functional consequences. Due to its prevalence and the morbidity it causes, diabetic foot has become a public health problem. Our retrospective cross-sectional study is primarily descriptive. Its aim is to determine the frequency of diabetic foot in the surgical department of the Panzi General Referral Hospital during the period from January 2017 to December 2021. During these five years, the hospital recorded 745 hospitalized patients, of whom 25 were included in this study. The frequency was 3.4%, the mean age was 46.84 +/- 22.4 years (range 3 to 80 years), and patients aged 41 to 60 years were the most affected, representing 44% of our series. There was a female predominance, with 68% of the cases studied, and a sex ratio of 0.47. The majority of patients came from urban areas (64% of cases). Arteriopathy was the most frequent type of diabetic foot (48%), with gangrenous lesions accounting for another 48%, and type 2 diabetes being the most prevalent (68%). Conversely, 94% of patients admitted to the surgical department of the Panzi General Referral Hospital were discharged with their lesions stabilized. Diabetic foot is common among diabetic patients in South Kivu. Combating this scurge requires patient and healthcare staff education, as well as multidisciplinary and coordinated care. Our study and literature review highlighted the epidemiology of diabetic foot in order to determine its prevalence in South Kivu. Keywords: Epidemiology, diabetes and diabetic foot.
- BeatMV
Turn any song into scroll-stopping videos in minutes with AI-directed scripts and
- Disclaimr — AI COI Compliance for SMBs
A COI is just a PDF — we verify it's actually real
- Sakana Fugu
Sakana Fugu
- Diaform
Diaform
- SuperIntern
SuperIntern
- remio: Your Personal ChatGPT
Get Tailored Answer with Your Personal ChatGPT
- Managed Hermes Agent | Hostinger
Hermes Agent with zero setup, maintenance, or infrastructure management.
- Toonly AI
Edit and generate images with AI prompts instantly.
- AirMusic - AI Music Video Generator
Turn any song into viral music videos.
- AI website & app builder | Hostinger
Launch your idea today
- Nvidia recruits for humanoid robotics roles in China
Nvidia's job openings include work on Project GR00T, a humanoid robot foundation model.
- APROPOS
your hire starts with the clarity of the role
- AnuCal
Most calendars manage weeks. AnuCal helps manage years.
- Unloop
Break the scroll loop
- Memorra
Smart planning for every celebration.
- VCPeer
Plain-English search for VC and private market data