搜索结果：coding

Artificial intelligence (AI), particularly generative AI and large language models, is increasingly used for assessment-related tasks in medical education. Existing overviews often address AI in medical education broadly, limiting assessment-specific interpretation of functions, settings, learner stages, and responsible-AI reporting domains. This study aims to map the literature on AI for assessment and feedback in medical education, including publication trends, bibliometric structure, assessment functions, AI types, settings, learner stages, and reporting of validity, reliability, fairness, integrity, transparency, human oversight, implementation, and governance. We conducted a bibliometric mapping study incorporating structured thematic evidence-map coding. Web of Science Core Collection, Scopus, and PubMed were searched from January 1, 2015, to April 8, 2026. Document selection and main evidence-map coding were based primarily on titles, abstracts, and bibliographic metadata, with targeted ambiguity resolution. Because reporting domains may appear mainly in full-text sections, all 435 included records underwent full-text sensitivity analysis for the 8 reporting domains. Coding reliability was assessed by two coders using percent agreement and Cohen κ before adjudication. Exploratory subgroup analyses and an excluding-2026 partial-year sensitivity analysis were conducted. Searches identified 14,968 records; 435 were included after deduplication and selection. Overall, 399 (91.7%) records were indexed in the post-ChatGPT period. Generative AI was coded in 310 (71.3%) records, and large language models in 301 (69.2%) records. In the assessment-function umbrella analysis, learner performance evaluation accounted for 270 (62.1%) records, feedback for 93 (21.4%) records, assessment content generation for 65 (14.9%) records, and other or unclear functions for 7 (1.6%) records. The most common settings were board-style examinations (n=151, 34.7%) and written examinations (n=88, 20.2%); undergraduate medical education was the most represented learner stage (n=172, 39.5%). Full-text-confirmed reporting was most frequent for reliability (n=288, 66.2%) and implementation (n=231, 53.1%); intermediate for validity (n=158, 36.3%), fairness (n=132, 30.3%), and transparency (n=130, 29.9%); and less frequent for governance (n=57, 13.1%), human oversight (n=46, 10.6%), and integrity (n=26, 6%). Stage 1 κ values ranged from 0.785 to 0.895, and stage 2 κ values ranged from 0.809 to 0.880. Excluding 65 partial-year 2026 records did not change the overall interpretation. The indexed English-language literature on AI for assessment and feedback in medical education expanded rapidly in the post-ChatGPT period and was concentrated in generative AI, large language models, examination-oriented assessment, and undergraduate medical education. Future studies should complement examination benchmarking with authentic assessment contexts, distinguish assessment content generation from learner-facing evaluation and feedback, and report responsible assessment domains more consistently.

To explore aspects of professional identity formation in participants of a longitudinal pediatric program for final-year medical students in the US. Many institutions have specialty-specific courses focused on the transition to residency, typically at the end of medical school. This study examines professional identity formation among participants of a longitudinal fourth-year Pediatric Concentration. Authors used a case study design for this qualitative research study using semi-structured interview questions based on a conceptual model of professional identity formation. Authors used purposeful sampling to identify study participants who were randomly selected from former students who completed the program between 2019 and 2021. Three interviewers independently conducted interviews with individual program participants. Thematic analysis with coder reliability was utilized for data analysis; two investigators independently coded and performed inter-rater reliability (IRR) on three sets of questions from transcriptions. After ensuring acceptable IRR, investigators coded the remaining questions and subsequently identified themes and sub-themes. Eleven former students completed interviews. Following initial independent coding, IRR for three questions was 80%. After coding the 11 interviews, investigators determined that code saturation was achieved. They subsequently identified six main themes: career development, interpersonal connection, personal growth, positive role models, skill-building, and supportive learning environment. Two cross-cutting themes were recognized throughout the data: sense of belonging and gaining confidence toward starting residency. Participants in a longitudinal specialty-specific program in the final year of medical school experienced professional identity formation through a supportive learning environment fostering connection and personal growth while building skills, surrounded by positive role models. This community of practice cultivated a sense of belonging and helped participants gain confidence toward their residency and career. This longitudinal, specialty-specific approach to the final year of medical school may enhance professional identity formation for students entering all specialties and could facilitate transition to internship.

Acute lymphoblastic leukaemia (ALL) is characterized by uncontrolled proliferation of lymphoid progenitor cells. Advances in genomic and epigenomic profiling have enabled the identification of over 40 molecular subtypes defined by distinct genetic drivers, transcriptional programmes and regulatory alterations. These insights have refined the classification, particularly of B cell precursor ALL (B-ALL) and are increasingly informing risk stratification, therapeutic decision-making and disease monitoring. By contrast, the classification of T cell ALL (T-ALL) has historically relied on immunophenotypic criteria, but recent large-scale genomic studies have uncovered biologically defined subtypes driven by diverse coding and noncoding alterations. Many genomic lesions represent clinically actionable vulnerabilities, including kinase-activating alterations that have enabled the use of targeted therapies. However, treatment resistance remains a major challenge, arising through clonal evolution, acquisition of secondary mutations and adaptive transcriptional and epigenetic reprogramming. In this Review, we highlight recent advances in understanding of the biological basis of ALL, with a focus on recently identified genetic alterations, gene expression patterns, alterations in three-dimensional genome architecture and epigenetic regulation that drive ALL initiation, progression and therapeutic response. Furthermore, we discuss how genetic heterogeneity contributes to clinical variability and how integrating molecular and biological insights can improve risk stratification and therapeutic outcomes.

Context-dependent decision-making enables flexible behavior by allowing identical sensory inputs to guide different actions depending on memory, rules, or goals. Recent advances in large-scale neural recordings have shifted the focus from single-neuron tuning to population-level representations, revealing principles by which neural populations support such flexibility. Here, we review evidence of how context-dependent decisions are implemented by population coding mechanisms, including nonlinear mixed selectivity, task-dependent population geometry, shared representational subspaces, and structured across-neuron correlations. Nonlinear mixed selectivity expands representational dimensionality, allowing downstream readout of arbitrary combinations of task variables. Learning reshapes population geometry, and it may balance flexibility and generalization by promoting the reuse of shared representations when task components overlap. Structured correlations between neurons that share a projection target enhance transmission of context-dependent information to downstream circuits. These population-level coding mechanisms provide a conceptual framework for understanding how neural circuits integrate sensory and contextual information to guide behavior.

The King Ratsnake (Elaphe carinata), a widely distributed non-colubrine snake in East Asia, exhibits a remarkable capacity for broad-spectrum venom resistance, enabling it to prey on other snakes, including lethal viperid and elapid species. Despite its ecological and physiological significance, the genetic basis of this trait remains largely unknown. Here, we present a chromosome-level genome assembly of E. carinata, generated using PacBio HiFi long-read sequencing, Illumina short-read data, and RNA-seq-supported by a synteny-based scaffolding approach using the closely related Elaphe schrenckii. The 1.62 Gb assembly (contig N50 = 2.77 Mb, scaffold N50 = 143.07 Mb) achieves 97.3% BUSCO completeness, with over 90% of sequences anchored into 18 pseudochromosomes. Repetitive sequences account for 53.19% of the genome, with LINE elements being the most abundant. We annotated 19,750 protein-coding genes, of which 99.2% were functionally assigned using integrated evidence from homology, transcriptomics, and ab initio predictions. This high-quality genomic resource provides a foundation for exploring colubrid evolution, venom resistance mechanisms, and the development of novel antivenom therapies.

In 2021, a total of 82 million people used electronic cigarettes (e-cigarettes) globally. E-cigarette regulations around the globe vary widely from the product being banned in some jurisdictions to being completely unregulated in others. The Tobacco Pack Surveillance System (TPackSS) was initiated in 2012 to monitor tobacco packs available in 14 low- and middle-income countries with the greatest number of people who smoke. The aim of TPackSS is to assess compliance with country-specific tobacco packaging and labeling requirements and identify marketing features and appeals used on tobacco packaging. The objective of this study was to adapt and expand the previous TPackSS protocol to also include disposable e-cigarette devices and their accompanying products: e-cigarette liquids and pods or cartridges. E-cigarettes were added to TPackSS data collection in Indonesia in 2022 and in China in 2023. Collection took place in Jakarta, Medan, and Surabaya in Indonesia and Shanghai, Beijing, Chongqing, Guangzhou, Kunming, and Shenzhen in China. A total of 15 neighborhoods per city in Indonesia and 12 per city in China were visited. The TPackSS protocol developed for tobacco products in 2012 to 2013 was used as the foundation for the adapted protocol for e-cigarettes. The adaptation of the original TPackSS protocol followed an iterative process involving extensive discussions among the TPackSS study team, comprising researchers with expertise in tobacco control and 13 years of experience in tobacco product packaging surveillance. The study adapted the sampling frame and sampling strategy, pack collection procedures, photography guide, shipment and translation procedures, and codebook to account for the unique country contexts for e-cigarettes (eg, regulations, market for e-cigarettes, and retailer landscape) and the heterogeneity of e-cigarette packaging observed. These adaptations were based on the review of peer-reviewed literature, white papers, Euromonitor data, and consultation with researchers with expertise in e-cigarette marketing and with in-country public health professionals. Using Indonesia as a pilot country, the protocol was evaluated and further refined post implementation for use in China. The TPackSS study began in 2012 and was funded by Bloomberg Philanthropies. Data collection took place from September to October 2022 in Indonesia and April to May 2023 in China. Across Indonesia and China, 968 unique e-cigarette products were collected. TPackSS data collection, coding procedures, and study findings are publicly available for use on the TPackSS website. The protocols presented here, which were used in 2 countries with contrasting e-cigarette markets and regulatory requirements, can be adapted for use in other countries. Findings from use of this adapted protocol can inform policy by providing insights into the design features and marketing appeals of e-cigarette products available on the market, as well as compliance with health warning label requirements where applicable.

Prostate Cancer Antigen 3, a type of long non-coding RNA, exhibits outstanding specificity as a biomarker for the diagnosis of prostate cancer, offering a highly effective diagnostic indicator, whereas the currently used prostate specific antigen exhibits low specificity, leading to reduced accuracy in prostate cancer diagnosis. Herein, we designed a novel fluorescent sensing platform for targeted detection of PCA3, which integrates the non-specific trans-cleavage activity of the CRISPR/Cas12a with the remarkable fluorescence quenching effect of Gold Nanorods. The Cas12a recognizes and binds to specific sequences of PCA3, thereby activating nonspecific cleavage activity, which cleaves fluorescent reporter probes adsorbed on AuNRs, thus leading to the recovery of fluorescence signals and enabling sensitive detection. The proposed fluorescent sensor exhibits excellent accuracy and convenience for the detection of PCA3 in urine, and a detection limit as low as 1.65 pM was obtained. This sensing system has achieved effective detection of clinical samples of prostate cancer and is expected to provide significant assistance in the screening and therapeutic feedback of prostate cancer in clinical diagnosis.

In May of 2022, an aquarium-maintained broadnose sevengill shark (Notorynchus cepedianus) developed proliferative skin lesions that prompted pathologic and molecular investigation. Histopathologic examination revealed epidermal hyperplasia consisting of proliferation of spinous epithelial cells with mild dysplasia. Metagenomic sequencing identified a novel adomavirus with an 18,834 base pair circular double-stranded DNA genome. The virus, provisionally named broadnose sevengill shark adomavirus (7AdoV), contains two bidirectionally expressed protein-coding gene sets. Genomic annotation and structural predictions of proteins were used to contextualize 7AdoV phylogenetically and functionally. Transcriptomic analysis showed that expression of the structural late gene set was higher than the replicative early gene set at the time of diagnostic sampling. In situ hybridization using RNAscope technology localized transcripts of the adomavirus Wasp gene to epithelial cells of the hyperplastic epidermis. Infection by this novel adomavirus was associated with superficial and proliferative lesions that were self-limiting and resolved in this shark.

Recent advocacy efforts have resulted in 49 out of 50 U.S. states passing some form of dyslexia legislation. This legislation may serve as a starting point for educators, clinicians, and parents to advocate for universal oral language screening as well as dyslexia screening. The purpose of this exploratory study is to investigate the degree to which oral language skills are already included in existing policies. We used deductive and inductive coding procedures to identify key words related to oral language skills. We then performed document analysis on 156 legislative documents related to dyslexia to describe the following: (a) to what extent states include "language disorder" or other related terms in their descriptions of disorders encompassed by the legislation; (b) to what extent legislation includes keywords related to oral language; and (c) how language keywords are distributed across screening, intervention, preservice preparation, and professional development requirements. Two states include "language disorder" in the scope of dyslexia legislation. An additional eight states have legislation with terms that could encompass a language disorder. Out of the 49 states with dyslexia legislation, 29 include at least one language-related keyword. The most common language keywords found in such legislation were "comprehension" and "vocabulary." Keywords were mentioned across screening, intervention, preservice preparation, and professional development requirements. In many states, existing legislation lays the groundwork for implementing universal oral language skill screening. Further legislation dedicated to universal oral language screening is also necessary to ensure children with language disorders are identified early.

Understanding complex relations between neuronal activity and animal behavior is central question in neuroscience. Rapid advancements in Artificial Intelligence (AI) methods offer powerful tools to investigate highly non-linear mapping between motor cortex activity and body movements. Here, we developed a Generative Adversarial Network (GAN) that showed that detailed videos of behaving rats can be recreated from activity of just few selected neurons. This analysis also revealed that the predictability of behavior from neuronal activity (and vice versa) initially increases as a rat learns a new task. However, after the animal performance on the motor task achieves the required accuracy, then coupling between neuronal activity and behavior decreases, without degrading task performance. A plausible interpretation is that, as training progresses from Early to Mid training days, more neurons become engaged, forming a denser, broadly distributed representation, which then in the Late training days evolves into a sparse and more energy-efficient representation, with only a small subset of tuned neurons. Neuronal network simulations showed that such changes in coding strategy may be explained by neurons minimizing their energy use. Thus, our approach reveals a non-linear relationship between learning stages and neural-behavioral coupling, which is likely driven by energy efficiency.

Intervertebral disc degeneration (IDD) is the leading pathological cause of low back pain, while current clinical treatments are only palliative and cannot reverse the programmed cellular senescence driven by epigenetic dysregulation. This process is characterized by progressive loss of nucleus pulposus (NP) cell identity and establishment of a self-amplifying senescence-associated microenvironment. In this review, we synthesize recent advances elucidating how heterogeneous senescent cell populations and their secretory phenotype (SASP) orchestrate a destructive vicious cycle in IDD. We further dissect the synergistic interplay among DNA methylation, histone modifications, and non-coding RNAs that constitutes the "epigenetic aging clock" and drives premature cellular aging within the disc. Notably, we evaluate emerging therapeutic strategies aimed at clock reversal, including senolytic clearance of senescent cells, epigenetic remodeling using small-molecule inhibitors or CRISPR-dCas9 editing, and cellular reprogramming approaches ranging from iPSC differentiation to direct lineage conversion. We propose a synergistic "clear, prime, then seed" roadmap that sequentially combines these interventions for optimal regeneration. This work provides a systematic theoretical framework for the clinical translation of epigenetic-targeted therapy for IDD, and breaks through the cognitive limitation of traditional mechanical wear theory.

Low-dose ionizing radiation (LDIR, ≤100 mGy) is a public health concern due to its extensive use in diagnostic and therapeutic imaging. This study examined somatic variants (SV) among Korean industrial radiographers exposed to LDIR using whole-genome sequencing (WGS). WGS data from 65 workers (mean age 36.7 years) collected between 2016 and 2023 were analyzed. Participants had a mean employment duration of 12.7 years and an average cumulative radiation dose of 33.9 mSv from National Dose Registry records. SV were identified via the GATK Mutect2 single-sample workflow applied to peripheral blood-derived DNA and sequentially filtered using standard pipelines, including FilterMutectCalls, population frequency, recurrent artifacts, Funcotator annotation, and manual Integrative Genomics Viewer review. A total of 105 170 somatic variants were identified, with a median of 1744 variants per individual, approximately 98% of which were single nucleotide variants. Cumulative radiation dose showed a significant positive correlation with total SV burden (R=0.32, P=0.013), including both coding (R=0.34, P=0.008) and non-coding regions (R=0.31, P=0.015). Age, smoking, alcohol consumption, hypertension, and hyperlipidemia were not significantly associated with variant burden. These WGS-based findings provide preliminary insight into blood-derived SV patterns among occupationally exposed radiation workers. Given the small sample size, blood-only design, and absence of matched unexposed controls, the findings should be interpreted as hypothesis-generating and require validation in larger longitudinal studies with comprehensive exposure assessment.

We recently reported that GADD34, an integrated stress response (ISR)-associated protein expressed at low levels in many cell lines, functions as a novel HIV-1 restriction factor. To elucidate the mechanism underlying GADD34-mediated inhibition of HIV replication, we investigated the role of long non-coding RNAs (lncRNAs). Here, we identify and characterize a novel lncRNA, GRHAL1 (GADD34-regulated HIV accessory lncRNA-1), which is highly expressed in GADD34-knockout (GADD34-KO) cells relative to WT cells. GRHAL1 is expressed at low basal levels in Jurkat and MT-2 CD4+ T cell lines and in primary human CD4+ T cells and is upregulated following HIV-1 infection. T cell activation signals that promote HIV-1 replication also induced GRHAL1 expression; however, this induction was independent of IFN and ISR signalling. Transfection of in vitro - synthesized GRHAL1 significantly enhanced HIV-1 LTR-driven gene expression from both integrated and unintegrated promoters. GRHAL1 stimulated both Tat-independent and Tat-dependent LTR activation, with a more pronounced effect observed at suboptimal Tat levels, indicating functional synergy between GRHAL1 and Tat under limiting Tat conditions. Mutational analysis of the LTR demonstrated that GRHAL1-mediated activation requires an intact Tat-responsive TAR element and upstream Sp1-binding sites, linking GRHAL1 activity to Tat and Sp1 in transcriptional regulation. This synergy was further confirmed using a minimal HIV-1 promoter containing Sp1 sites and a TAR sequence. Moreover, GRHAL1 alone was sufficient to activate a heterologous promoter containing Sp1-binding elements. The Sp1-selective inhibitor mithramycin suppressed both Tat-independent LTR activation and the cooperative activation mediated by GRHAL1 and Tat. Electrophoretic mobility shift assays, RNA immunoprecipitation and RNA-pulldown experiments demonstrated that GRHAL1 directly interacts with Sp1 and that Tat enhances this interaction, suggesting that GRHAL1 modulates Sp1 activity through direct binding. In summary, we identify GRHAL1 as a novel HIV-1-induced lncRNA that regulates viral gene expression by interacting with Sp1 and synergizing with Tat to enhance HIV-1 transcription.