Search Results (3,216)

Subtitles often attract visual attention even when they are not necessary for comprehension. In the present eye-tracking experiment, we examined whether attention to subtitles in instructional videos varies as a function of audio–subtitle language–script pairing in Hindi–English bilinguals with an English-medium instruction (EMI) background. Native Hindi participants viewed videos in three conditions: English audio with English subtitles (L2–L2), Hindi audio with Hindi subtitles (L1–L1), and English audio with Hindi subtitles (L2–L1). In the L2–L2 condition, gaze was distributed similarly across speakers’ faces and subtitles. In contrast, in both Hindi-subtitle formats, viewers allocated more dwell time to the speakers’ faces than to the subtitles. Comprehension scores did not differ significantly across conditions. These findings suggest that subtitle engagement among EMI bilinguals is not solely determined by the presence of subtitles but is also modulated by the properties and perceived utility of the written channel. More generally, our results caution against the view that subtitle engagement is uniformly automatic across multilingual instructional settings. Full article

The Five Domains Model has become one of the most widely adopted frameworks in animal welfare science and practice. The Model is now applied in a range of ways; among the most prominent are (1) as a framework for systematic and structured welfare assessment and (2) as an organising structure for planning and communicating appropriate (i.e., welfare-focused) care provisions, education, and standards. This paper focuses on these two applications and proposes a corresponding update to the affiliated Five Provisions and Welfare Aims. Specifically, we revise: (1) Provision 4 from “Appropriate Behaviour” to “Appropriate Choices” to reflect the 2020 update of the Model incorporating human–animal interactions and the 2023 operationalisation of agency in Domain 4; (2) Provision 2 from “Good Environment” to “Good Living Space” to resolve ambiguity with Domain 4’s “Interactions with the Environment”; and (3) Provision 5 from “Positive Mental Experiences” to “Integrated Care,” which captures consistent delivery of the first four provisions over time and across all those who interact with the animal. This update also pairs Provision 5 with a welfare aim that specifies the integrated mental state the animal should experience as a result. This change makes the distinction between care (provisions) and welfare (aims) consistent throughout the framework. It also makes explicit the integrative role of Provision 5, which parallels Domain 5’s role in the Model. We then describe the reasoning process that distinguishes welfare assessment from welfare-focused care provision. Welfare assessment uses the domain structure as a reasoning pathway, with the assessor using indicators and their impacts in Domains 1 to 4 to infer named mental (affective) experiences in Domain 5. Planning and communicating appropriate (i.e., welfare-focused) care uses the same structure to organise information about what is provided to animals, without executing the inferential step to Domain 5. Drawing on examples from organisations that use the Model for different purposes, we show that both applications are legitimate but produce different outputs. The Five Provisions framework, with its dual structure of provisions paired with welfare aims, serves the care planning and communication function more effectively than does the Model’s domain structure alone. Recognising these different uses also helps to locate where recent critiques of the Model apply and where they do not. Finally, we propose that the provisions and welfare aims framework can supplement “needs” language in legislation and policy to better reflect the distinction between animal care and animal welfare. Full article

Background: Mild Cognitive Impairment (MCI) is a clinical state between normal aging and dementia. It may involve impairment in one or several cognitive domains. MCI offers a key window for maintaining cognitive function and studying how deficits develop in the elderly, making it of great research value. Measurement tools for screening MCI are not yet standardized in China. The accuracy of diagnostic criteria and threshold values needs improvement. Previous studies on the neural mechanisms of MCI have examined various aspects, but the changes in the white matter microstructure in older adults with MCI remain unclear. Most past studies used Fractional Anisotropy (FA) analysis to examine changes in white matter fiber orientation, often ignoring fiber density. As a result, findings are often contradictory or difficult to interpret. Therefore, it is necessary to assess cognitive function in MCI populations using more comprehensive and standardized measurement tools. It is also important to explore the association between changes in white matter microstructure and cognitive function in MCI by analyzing FA and Mean Diffusivity (MD). Methods: First, we assessed cognitive function using the Cognitive Function Measurement Scale for the Elderly, developed by Beijing Normal University, with diagnoses based on the NIA-AA (National Institute on Aging—Alzheimer’s Association) criteria. Second, we employed Diffusion Tensor Imaging (DTI) combined with Tract-Based Spatial Statistics (TBSS) to investigate alterations in the white matter fiber tract integrity in individuals with MCI. Based on the metrics used, this study was divided into two analytical approaches: Analysis Mode 1 utilized FA to explore changes in white matter fiber orientation in the MCI group. Analysis Mode 2 utilized MD to examine changes in white matter fiber density in the MCI group. Third, we further explored the association between alterations in the white matter fiber tract integrity and cognitive function in individuals with MCI. Specifically, FA and MD values from brain regions showing significant differences between the MCI and normal control groups were extracted and correlated with cognitive test scores. Results: According to the results of the community measurement survey, the prevalence of MCI among the elderly in Shenzhen is approximately 21.54%. Individuals with MCI exhibited functional decline in memory, attention, language, executive function, and spatial processing. DTI results indicated that (1) FA values across the brain’s white matter fiber tracts showed a decreasing trend in the elderly with MCI, with no areas exhibiting significantly higher FA values. Specifically, FA values were significantly lower in the corpus callosum, internal capsule, corona radiata, thalamic radiation, external capsule, superior fronto-occipital fasciculus, and cingulum (cingulate gyrus). (2) White matter fiber tracts with significantly reduced FA values also demonstrated significantly increased MD values. Additionally, MD values in the cingulum (hippocampus), inferior cerebellar peduncle, and corticospinal tract were significantly reduced in the MCI group. (3) Correlation analysis revealed that the significant differences in FA and MD values within the white matter fiber tracts of older adults with MCI were correlated with scores on several cognitive tests. Conclusions: In the present study, older adults with MCI tended to exhibit functional decline across multiple cognitive domains and relatively extensive microstructural white matter damage. Observations suggested that white matter fiber density may be informative regarding these microstructural alterations, indicating that diffusion biomarkers in key regions such as the cingulum (hippocampus) warrant further investigation. Full article

Generative artificial intelligence (AI) is transforming biological and medical research and data analysis. Beyond analyzing existing information, these models can learn complex patterns and generate new data such as realistic protein sequences, genetic variants, or clinical notes. In molecular biology, language-like sequence models can read and generate DNA, RNA, and amino acid sequences to predict genetic variant effects, design new proteins, and explore molecular functions. In medicine, large language models (LLMs) trained on biomedical literature and electronic health records (EHRs) can summarize clinical findings, identify patterns, and provide decision support for clinicians and healthcare providers. Additionally, synthetic data generation can help protect patient privacy and augment existing disease datasets. While these advances make tasks that were previously impractical possible at scale, they also carry major risks, including producing convincing but incorrect results, reflecting hidden biases in the training data, and underperforming when real-world conditions change. Full article

Background/Objectives: Higher education requires students to communicate in complex academic and social contexts, including oral presentations, group work, help-seeking, assessment, and peer interaction. For students with developmental communication disorders, and communication-related developmental profiles, these demands may create hidden participation vulnerabilities that affect mental health, academic engagement, and belonging. This scoping review mapped empirical evidence among tertiary students, focusing on mental health, academic participation, social belonging, institutional support, and contextual influences. Methods: A scoping review was conducted in accordance with PRISMA-ScR guidance. Five databases, PubMed, PsycINFO, CINAHL, Scopus, and Web of Science, were searched for English-language, peer-reviewed empirical studies published from 2000 onwards. Eligible studies involved university, college, or tertiary students with developmental speech, language, fluency, pragmatic communication, or communication-related developmental profiles, who reported at least one mental health, academic, or social participation outcome. Data were charted and synthesised thematically, with methodological quality appraised using CASP-informed criteria. Results: Twenty-one studies were included. Evidence was strongest for stuttering and fluency-related participation, while research on developmental language disorder, speech sound disorder, pragmatic language impairment, cluttering, and mixed communication profiles was limited. Across studies, communication needs intersected with anxiety, depression, stress, self-efficacy, oral assessment, help-seeking, disclosure, stigma, accommodation access, and belonging. Key limitations included reliance on self-report, cross-sectional or retrospective designs, inconsistent diagnostic confirmation, and limited evidence for intervention. Conclusions: The available evidence suggests that developmental communication disorders and communication-related developmental profiles can function as hidden participation vulnerabilities in higher education. These vulnerabilities are shaped by students’ communication profiles and by communication-intensive university environments. Universities may therefore need communication-accessible teaching, flexible assessment, visible support pathways, and coordinated support across disability services, counselling, academic support, and speech–language pathology. Full article

Genetic programming (GP) has been widely used in quantitative finance for discovering formulaic alpha factors that can predict asset returns. However, GP often produces overgrown expressions that are difficult to interpret and expensive to evaluate. This paper proposes a large language model (LLM)-assisted pruning framework that reviews expression trees generated by GP, with the LLM acting as a semantic reviewer that flags redundant or financially implausible branches based on structural complexity and contextual reasoning. The proposed method is formalized as a closed-loop Trigger–Evaluate–Decide–Execute (TEDE) process. We present mathematical formulations, algorithmic design, and examples showing how redundant nested functions can be simplified while monitoring predictive performance. Experiments with high-frequency cryptocurrency market data, using DeepSeek-V4-Flash as the semantic engine, show lower expression complexity and higher rubric-based interpretability scores for the pruned symbolic factors. Under the reported test setup, the LLM-pruned configuration has higher Information Ratio (IR) values than the listed baselines and more compact expression trees than the GP baselines. Full article

We address the paradoxical transformation of a classical-mechanical particle motion when the space and time scales of observation pass below the uncertainty principle threshold. This is analyzed in the language of classical statistical mechanics, considering specifically many-particle systems inhomogeneous along one spatial direction. We employ the density functional formalism in its square-gradient form and find: (i) The macroscopic solution is analogous to the classical trajectory of a particle under a potential of force given by (minus) the free energy density. Whereas, (ii) fluctuations around the solution in (i) are equal to the quantum-mechanical wave functions of a particle under a potential given by the curvature of the free energy density. We illustrate this situation with three textbook examples: A particle in a box, the harmonic oscillator, and the hydrogen atom. We show that their time-independent Schrödinger equation wave functions describe, respectively, the fluctuations of a fluid interface, of critical point fluctuations, and of a confined ideal gas. At large scales, sharp probability distributions make fluctuations irrelevant; the vanishing of the first variation yields the macroscopically observable statistical-mechanical non-uniformity, equivalent to the classical particle trajectory. But at sufficiently small scales, with necessarily very few particles, distributions appear much wider, fluctuations dominate, and one obtains the Schrödinger equation (for the microscopic potential). Full article

The bilingual advantage hypothesis proposes that bilingual children will display greater executive function (EF) skills compared to their monolingual peers. However, most research on this topic neglects to include monolingual children from both language groups for comparison, thus confounding language status and cultural context. To address this gap, we administered an extensive battery of EF tasks to 189 typically developing children ages 47–95 months (M_age = 71.47, SD = 11.68, 42.9 % Female) drawn from three language status groups: Korean-English Bilingual and English Monolingual (both in the northwestern United States) and Korean Monolingual (South Korea). Korean-English Bilingual children scored significantly higher on the EF composite than Korean Monolingual children, even after controlling for child age and verbal ability. Both English Monolingual and Korean-English Bilingual children waited significantly longer during a delay-of-gratification task than Korean Monolingual children when controlling for age and verbal ability. Korean-English Bilingual children outperformed English Monolingual and Korean Monolingual children on the Comprehensive Test of Nonverbal Intelligence. There were no significant differences between language status groups on the other individual EF tasks after adjusting for multiple comparisons. Taken together, we did not find consistent support for a bilingual advantage in EF skills: Country of residence also played a role, with children living in the United States outperforming children living in Korea in some cases. Full article

Background/Objectives: Lung cancer is a leading cause of cancer-related mortality worldwide. As patients increasingly utilize large language models (LLMs) for health information, evaluating the readability and patient-centeredness of these tools is critical. This study aims to compare the performance of ChatGPT-4o mini, Microsoft Copilot, and Google Gemini in providing lung cancer information, focusing on their utility for individuals with limited health literacy. Methods: In this cross-sectional study (March 2026), 30 responses to ten standardized lung cancer-related queries were analyzed. Outputs were assessed using JAMA benchmarks and mDISCERN for quality, the SMOG index for readability, and PEMAT-P for understandability and actionability. Inter-rater reliability was analyzed using intraclass correlation coefficients (ICCs). Results: ChatGPT-4o mini demonstrated superior readability, achieving a sixth-grade level (SMOG: 6.23 ± 0.72, p < 0.001). Gemini achieved higher JAMA scores, indicating stronger academic rigour. While PEMAT-P scores were highest for ChatGPT (63.7%), all models exhibited moderate mDISCERN quality. Inter-rater reliability was excellent for JAMA (ICC = 1.000) and PEMAT-P (ICC = 0.883), though moderate for mDISCERN (ICC = 0.365), reflecting inherent interpretative subjectivity in qualitative assessment. No hallucinations were observed. Conclusions: Current LLMs exhibit a trade-off between accessibility and academic rigour: ChatGPT favours patient-friendly readability, while Gemini emphasizes structured content. The observed inter-rater variability in mDISCERN underscores the complexity of standardizing qualitative AI evaluation. These findings suggest that LLMs function best as complementary aids rather than substitutes for physician-led communication. Full article

Reasoning-intensive multimodal retrieval suffers from a counter-intuitive bottleneck: on MM-BRIGHT multimodal-to-text (Query+Image → Documents), the strongest dense multimodal encoder reaches only 27.6 nDCG@10 and the rest of the dense vision–language retrievers cluster between 10.0 and 23.0. The visual signal, encoded as a dense vector, adds noise rather than evidence; even augmenting strong text retrievers with raw image captions degrades performance by up to 12.0 points. We propose VISA, a Visual Symbolic Agent that re-casts multimodal-to-text as text retrieval over three parallel streams. A Vision LLM is dispatched in three roles via separate prompts: a zero-shot router that classifies the query image into up to three parser types from a fixed taxonomy of nine (chart, circuit, equation, screenshot, code, figure, diagram, map, photograph); typed parsers that extract structured text per type; and a holistic captioner. The agent constructs three text streams (raw query, query ⊕ symbolic, query ⊕ caption), scores each with a single frozen 4B-parameter retrieval LLM, and fuses the per-document scores via Reciprocal Rank Fusion or a confidence-weighted linear combination. The whole agent contains no trainable parameters. The key novelty is a change of substrate: rather than projecting the query image into a dense multimodal vector that competes with text, VISA is, to our knowledge, the first retrieval system to convert the image into typed symbolic text and keep retrieval entirely text-side, so that a frozen text retriever can match the literal tokens (axis values, variable names, function signatures) that answering documents actually contain. Across all 29 MM-BRIGHT multimodal-to-text domains, VISA achieves 32.4 nDCG@10, an absolute improvement of +4.8 over the strongest dense multimodal encoder and substantially larger margins over the remaining six dense vision–language baselines. Per-domain analysis shows VISA maintains its margin across STEM and software domains where image content is structure-heavy. In practical terms, VISA is training-free and model-agnostic: it requires no fine-tuning, reuses any off-the-shelf vision LLM and text retriever, caches all per-image parsing so re-runs cost only three query encodes, and can therefore be dropped into an existing text-retrieval stack to add reasoning-intensive multimodal capability without building or training a multimodal encoder. Full article

Large language models (LLMs) are increasingly used in programming education to provide on-demand conceptual clarification, yet how students actually use this feature in mastery learning systems (in which learners must demonstrate conceptual competence before progressing)—and whether clarification interactions relate to subsequent learning—has received limited empirical study. This paper analyzes 732 student remediation episodes (366 students, 43 assignments) to examine how students move through the remediation branch of an LLM-assisted programming course, whether their behavioral pathway choices are associated with subsequent coding challenge efficiency, and what theoretical role the clarification function plays. The results show that 78.0% of remediation episodes follow a pure retesting strategy, with only 22.0% involving any clarification interaction. Clarification is highly concentrated on conceptual questions (84.7%) and occurs mostly in the first remediation round (86.3%). An effect size analysis reveals a large difference in remediation rounds between single immediate and single delayed clarifiers (Cliff’s $\delta =-0.912$), suggesting that the timing of clarification is more strongly associated with remediation efficiency than its occurrence alone. mixed-effect linear models show no significant pathway effects on coding challenge process efficiency (active time and number of code snapshots; all $p>0.05$), a null result that is further examined through code-variability subgroup analyses. We argue that the clarification feature acts as a selective process-support mechanism: its observable value appears to lie in a shorter remediation process rather than in improved subsequent task efficiency, and this association is clearest when clarification occurs early. The findings have practical implications for the design of clarification features in AI-assisted learning systems and for instructional intervention strategies. Full article

Lifecycle management of the conversational AI agent, in the case of Copilot Studio and Dataverse as enabling technologies, is discussed in this paper. After an in-depth examination of the academic literature, policy reports, and lifecycle models, the research also concludes that there are AI applications to be utilized in the government sector, and there are policies to be revised, alongside some ethical considerations that can and must be implemented. It has also revealed that conversational AI is so on trend that governments are employing this technology to do even more, to socialize with and serve the needs of more people in multiple languages. They can also decrease response times by 40%. But its initial condition will not endure for long. Lifecycle continuous monitoring as well as lifecycle ethics and participative design should be practiced in lifecycle governance so that nobody feels sidelined, left without influence, or interrogated. Copilot Studio is a low-code or no-code orchestration environment that runs on your code, and Dataverse ensures your data will be compatible with other systems. In the study, the theory attempts to touch on the harmonization of entities, citizen security and technical functions in the lifecycle. In this model, we will differentiate why a field-conversational AI model would lead to the creation of a vibrant, responsible, and effective service model. The technical and ethical lifecycle management of the AI integration offers a structure of accountability in which governments should extend the conversational agent according to the values held by the government. Full article

Alzheimer’s disease and related dementias (ADRD) are neurodegenerative conditions characterized by progressive cognitive and functional decline. AD pathology is associated with extracellular amyloid-β plaques, intracellular tau neurofibrillary tangles, synaptic dysfunction, and neuronal loss. AD accounts for approximately 60–80% of dementia cases globally. In 2022, AD was the seventh leading cause of death in the United States, and the number of Americans aged 65 and older living with Alzheimer’s dementia is projected to increase substantially by 2060. Despite decades of research, AD/ADRD data resources remain fragmented across clinical, imaging, genetic, genomic, and therapeutic domains. This paper addresses that gap by providing a centralized review of widely used AD/ADRD databases and computational methods. We first summarize computational approaches used to analyze these datasets, including machine learning (ML), natural language processing (NLP), and biomedical imaging. We then review eight databases classified into three categories: Clinical and Population Data, Genetics and Genomics, and Drug Discovery and Therapeutics. Finally, we discuss real-world applications, including early diagnosis, clinical decision support, personalized medicine, and drug-mechanism analysis. This review identifies opportunities for future work in data harmonization, cross-database compatibility, and robust, generalizable AI models for AD/ADRD research. Full article

This study examines the psychometric functioning of the Basic Learning Ability Assessment (BLAA), a high-stakes placement assessment used in Taiwan’s Adaptive Guidance Placement System (AGPS) for junior high school students with intellectual disabilities (IDs). The sample comprised 203 ninth-grade students with ID from 47 public junior high schools in Taiwan, all of whom completed three operational multiple-choice forms of the BLAA. Using classical test theory (CTT), we examined item difficulty using proportion-correct indices, item discrimination using upper–lower group discrimination indices, distractor functioning by comparing response patterns between higher- and lower-performing examinees, and internal consistency reliability using the Kuder–Richardson Formula 20 (KR-20). The results show that most items fell within the average-to-easy range and demonstrated acceptable to strong discrimination. Distractor functioning was generally satisfactory, with most items containing no nonfunctioning distractors. KR-20 coefficients ranged from 0.904 to 0.926, indicating high internal consistency within each form. Functional Language and Social Adaptation showed relatively stable psychometric patterns, whereas Mathematical Skills displayed greater variability in item difficulty, discrimination, and distractor functioning. Overall, the findings provide initial CTT-based internal psychometric evidence regarding the item functioning and form-level reliability of the BLAA, while highlighting the need for ongoing item refinement, particularly in the Mathematical Skills domain. Full article

Cancer-related cognitive impairment (CRCI), also known as chemobrain or chemofog, is characterized by subjective and/or objective changes in attention, executive functions, memory, and processing speed in patients with non-CNS cancers, particularly women with breast cancer. This structured narrative review synthesizes current evidence on mechanisms, neuropsychological assessment, neuroimaging correlates, clinical and demographic risk factors, emerging artificial intelligence and machine learning applications, and non-pharmacological approaches to CRCI in breast cancer. A structured literature search was conducted using PubMed/MEDLINE, PsycInfo, and Clinical Key up to May 2026, with emphasis on studies published between 2023 and 2026. Peer-reviewed English-language studies involving adult breast cancer populations and addressing predefined thematic domains of CRCI were considered. Given the heterogeneity of study designs, assessment tools, interventions, and outcomes, the findings were synthesized narratively. Current evidence supports a multifactorial model of CRCI involving neurobiological, treatment-related, psychological, and behavioral mechanisms. Neuroinflammation, endocrine disruption, oxidative stress, glial alterations, and structural or functional brain changes may contribute to cognitive symptoms; however, the strength of evidence varies, and many findings remain correlational or preclinical. Non-pharmacological interventions, including cognitive training, physical activity, mindfulness-based and psychological approaches, and multimodal digital programs, appear promising as supportive strategies. However, evidence remains heterogeneous, with benefits more consistently reported for patient-reported outcomes, fatigue, emotional distress, and quality of life than for objective neuropsychological performance. CRCI in breast cancer should be approached as a heterogeneous condition requiring early recognition, standardized assessment, and multidisciplinary supportive care. Future research should prioritize longitudinal designs, harmonized endpoints, and a clearer distinction between subjective and objective outcomes. Full article