ABSTRACT This study investigates how university students’ use of argumentation shapes their behavioural and cognitive engagement with generative AI-mediated learning tasks. Using content analysis and quantitative analysis, we analysed interaction log data and final essay assignments from 12 undergraduate students over the course of a semester. Findings reveal that prompts with higher argumentation complexity, particularly those integrating claims, reasons, and evidence, elicited more reflective post-interaction behaviours, such as editing or stopping AI-generated content. These complex prompts were also associated with a greater quantity of prompt-response cycles within shorter timeframes, indicating more efficient and intentional engagement. Cognitively, both semantic similarity between AI responses and students’ final essays and the overall essay quality were shaped by specific argumentation elements, with reasoning playing a particularly influential role. Notably, longer interaction duration and greater prompt quantity improved essay quality but did not enhance semantic alignment with AI responses, underscoring the multifaceted nature of cognitive engagement. These findings contribute to understanding the nuanced dynamics of student-AI engagement, highlighting the need for tailored strategies to scaffold argumentation skills for meaningful behavioural interactions and deeper cognitive engagement with AI-mediated learning tasks.

ABSTRACT Educators in higher education face persistent challenges in scaling AI literacy across disciplines and helping novice learners understand abstract AI concepts. Although research on game‐based learning (GBL) reports mixed outcomes, few studies have examined its large‐scale use in mandatory, asynchronous AI literacy courses for diverse undergraduate populations. Addressing this gap, this study investigates a scalable GBL‐based AI literacy course delivered to 4898 first‐year undergraduates across disciplines. Using a mixed‐methods design with 311 valid pre‐ and post‐survey responses and 20 interviews, the study evaluates students' cognitive, behavioural, affective, and ethical learning of AI. Quantitative results show significant improvements in overall AI literacy across cognitive, behavioural, and affective dimensions, while ethical learning gains were not statistically significant. Qualitative findings suggest that GBL stimulated students' epistemic curiosity and engagement with AI ethics while revealing pedagogical, technical, and learner‐centered challenges. The study provides large‐scale empirical evidence and proposes an instructional design framework for scalable AI literacy integration in higher education.

How does myelin develop in human visual cortex? By combining immunohistochemistry with in vivo and postmortem magnetic resonance imaging of longitudinal relaxation rate (R 1 ), which increases with myelin content, we find that myelin and R 1 increase across development but follow distinct trajectories. Immunohistochemistry reveals two phases of myelination: an infant phase of limited oligodendrogenesis, with myelin restricted to deep cortical layers, followed by widespread myelination across all layers during childhood. Cortical R 1 also increases across development and correlates with myelin by childhood. However, in infancy, R 1 increases outpace myelin growth and instead tracks dendritic arborization, indicating that the microstructural drivers of R 1 change across development. We hypothesize that deep layer myelination in infancy contributes to early visual function whereas later myelination of superficial layers enables prolonged cortical plasticity and learning of complex visual behaviors.

Prenatal alcohol exposure (PAE) causes fetal alcohol spectrum disorders (FASDs), which are neurodevelopmental conditions characterized by behavioral dysregulation, learning deficits, and cognitive inflexibilities. Alcohol exposure is harmful at all stages of human gestation, including the third trimester. This developmental window, characterized by rapid brain growth, myelination, and neural circuit formation, may be particularly vulnerable, yet the long-lasting behavioral and sensory consequences of exposure during this period remain poorly understood. In this study, neonatal mouse pups were exposed to ethanol (EtOH) or Air vapor from postnatal day (P) 4 to P8, which is equivalent to a third-trimester alcohol exposure (TTAE) in humans. Blood ethanol concentrations measured at P8 reached approximately 250 mg/dL, consistent with binge-level exposure. Air- and EtOH-exposed mice were then assessed as adults at 5-6 months of age for locomotor activity, anxiety-related risky behaviors, recognition memory, and increased susceptibility to peripheral neuropathy, as indicated by sensitization to light touch following minor chronic constriction injury (mCCI) of the sciatic nerve. We found that TTAE was sufficient to produce long-lasting behavioral outcomes in a sex-dependent manner. Notably, EtOH-exposed males exhibited increased spontaneous locomotor activity and risky behavior, whereas EtOH-exposed females showed minimal or decreased changes compared to their respective controls. However, both EtOH-exposed male and female mice exhibited marked increases in light-touch sensitization, referred to as mechanical allodynia, following mCCI, a response absent in air-exposed controls. Together, these findings reveal that TTAE is highly detrimental to behavioral regulation and creates a vulnerability to developing neuropathic pain in adulthood.

Mesolimbic dopamine (DA) neurons are central to cue-guided reward seeking and action sequence learning. Yet, the mechanisms by which cue-induced DA neural activity drives goal-directed or habitual sequence execution remain unknown. We designed two novel tasks to isolate the effect of sequence-delineating cues on DA-driven behavioral strategies and learning. In the lever insertion fixed-ratio 5 task (LI5), the lever insertion marked sequence initiation. In the lever retraction fixed-ratio 5 task (LR5), the lever retraction served as both sequence termination and reward-predictive cue. We found that sequence initiation and termination cues differentially affect reward expectation during action sequences, with only the termination cue contributing to greater outcome devaluation insensitivity, automaticity and behavioral chunking. Mesolimbic fiber photometry recording revealed that this habit-like behavior was associated with a rapid backpropagation in DA signals from the reward to the immediately preceding cue and with attenuated DA reward prediction error signals, which reflected greater behavioral inflexibility. Finally, in absence of external cues, brief optogenetic stimulation of VTA DA neurons at sequence termination was sufficient to drive automaticity and, to some extent, behavioral chunking. Our results highlight the critical role of cue-evoked DA signals at sequence termination in driving the development of automated, habit-like sequence execution.

Neurobiological Signatures of Dyadic Transmission of Fear Extinction in Adolescent Trauma Exposure and Posttraumatic Stress.

Purpose This study aims to investigate how sustainability competency and team diversity influence students’ intention to promote Sustainable Development Goals (SDGs). It also explores the moderating effect of team diversity on the relationship between sustainability competency and intention to promote SDGs (SDG Intention). Design/methodology/approach Empirical data were collected through a questionnaire-based research methodology involving 20,501 students actively participating in the Kampus Mengajar Program. This study used purposive sampling, which is explained by the theory of planned behavior (TPB) and experiential learning theory (ELT). Findings This study found that sustainability competencies, including leadership, critical thinking, collaboration and adaptability, significantly influenced students’ intention to promote the SDGs. In addition, team diversity in educational background significantly moderated the relationship between sustainability competencies and intention to promote the SDGs. Practical implications The study shows that developing a curriculum that supports sustainability education, comprehensive student training and diverse team building is essential. Institutional support and collaboration between institutions are also needed to ensure the success of programs supporting SDG achievement. Originality/value The originality of this study lies in the innovative use of an integrated theoretical framework, specifically ELT and TPB, to examine how sustainability competencies can enhance students’ intention to promote the SDGs. This research advocates a diverse educational approach that fosters critical thinking, systemic analysis and proactive engagement, essential to addressing complex sustainability challenges. Furthermore, exploring the moderating effects of team diversity introduces a new dimension to the discourse, highlighting the enriching potential of diverse educational environments in enhancing collaborative skills.

With the rapid advancement of Generative Artificial Intelligence (GAI) technologies, their integration into higher education is becoming increasingly widespread. This transformation is not only reshaping students' learning approaches but also redefining the collaborative dynamics between humans and AI. Based on the triadic framework of Exploration-Exploitation-Adaptation, this empirical study (207 valid questionnaires from Chinese university students, analyzed via structural equation modeling) investigates the behavioral mechanisms and pathways influencing learning outcomes among university students engaged in GAI-assisted learning. It examines how role adaptation, self-efficacy, task-technology fit, and institutional support affect learners' exploration and exploitation behaviors, and how these behaviors in turn impact learning effect. The findings reveal that role adaptation and self-efficacy are the primary internal drivers of GAI-related learning behaviors, while institutional support and task-technology fit serve as essential external enablers. Both exploration and exploitation behaviors significantly enhance learning outcomes, with exploitation showing a more pronounced effect. The model demonstrates good fit and significant path relationships among variables. While the results are consistent with the proposed adaptation-exploration/exploitation-effectiveness pathway, they only reflect correlational evidence and do not establish a causal mechanism. Theoretically, this study enriches insights into human-AI collaboration in higher education. Practically, it offers guidance for the optimization of intelligent educational systems and the design of behavior-guided strategies.

α-Synucleinopathies are neurodegenerative disorders characterized by the aggregation and propagation of misfolded α-synuclein. In Parkinson's disease (PD), the most common α-synucleinopathy, the progression of motor and nonmotor deficits, and dopaminergic neuron loss, are closely linked to the spreading of misfolded α-synuclein. Pioglitazone, a PPARγ agonist, has shown neuroprotective effects in preclinical models. However, its dosing and its effects on behaviour and neuropathology have not been explored in a model of α-synucleinopathy. Here, we evaluate the safety and neuroprotective effects of four pioglitazone treatment regimens in an α-synucleinopathy model induced by a single supranigral administration of β-sitosterol β-D-glucoside (BSSG), which replicates motor and non-motor symptoms and dopaminergic neurodegeneration. Adult male Wistar rats were randomly assigned to untreated, mock, α-synucleinopathy (αSN), and four pioglitazone treatment groups with distinct dosing schedules. Behavioural assessments, including sensorimotor, motor performance, anxiety-like behaviour, learning, and memory, were conducted at three time points over a 31-day protocol. Histological analyses included quantification of TH+ dopaminergic neurons, Nissl-stained viable neurons, and α-synuclein aggregates in the substantia nigra pars compacta (SNpc). Administration of pioglitazone was well tolerated. The pioglitazone regimen, which combines prophylactic and twice-weekly postinduction doses, most effectively attenuated behavioural deficits. Histologically, a significant reduction in α-synuclein aggregates and preservation of dopaminergic phenotype were demonstrated in the SNpc. Pioglitazone exhibited potential neuroprotective effects in a progressive αSN model, particularly under a prophylactic plus sustained treatment scheme. These findings support pioglitazone as a promising modifying therapy for PD and α-synucleinopathies.

Digital affordances refer to the possibilities provided by digital environments for learners. In the context of nursing education, artificial intelligence (AI) chatbots currently offer multimodal learning approaches and demonstrate various possibilities for digital actions. Therefore, exploring the digital affordances of AI chatbots in nursing education is crucial for the continuous advancement of the field. To evaluate the digital affordances of AI chatbots in nursing education, focusing on the relationship between digital affordances and learning gains. We employed affordance theory to conceptualize the potential actions of AI chatbots and utilized a taxonomy of affective, behavioral and cognitive learning gains to conduct a systematic review in nursing education. A total of 25 studies were identified in this systematic review. The geographical distribution of the studies is mainly in Asia. The most used study designs were quantitative designs (n = 12) with sample sizes between 16 and 457. The duration of these studies is usually short, ranging from a few hours to 3 months. The included studies reported several digital affordances of AI chatbots in nursing education, including assistance provision, personalization, human-like conversing, distilling information, and fostering familiarity. However, four digital affordances-facilitation, enriching information, context identification, and ensuring privacy-still lack empirical support. The evidence for the digital affordances of AI chatbots in nursing education was dominated by cognitive learning gains (such as learning achievement, critical thinking, and problem solving) and followed by affective (such as learning interest, self-efficacy, and enjoyment) and behavioral learning gains (such as engagement, diagnostic skills and clinical practice). However, several studies reported no statistically significant improvement in certain cognitive learning gains, particularly knowledge acquisition and clinical reasoning competency. Similarly, limited evidence was found for improvements in learners' confidence and satisfaction. These findings suggest that the current evidence remains inconclusive. Future research should employ longer study durations and larger sample sizes to further examine the educational impact of AI chatbots.

Predicting student academic performance has become increasingly vital in the field of educational data mining, as institutions seek data-driven strategies to enhance learning outcomes. However, many existing models rely solely on behavioral indicators or static features, often overlooking the role of time and context in shaping learning behavior. This limitation reduces predictive accuracy and adaptability in academic environments. To address this challenge, this study introduces EduFuseNet, a hybrid deep learning framework that integrates behavioral and spatiotemporal data for accurate classification of student performance. The workflow begins with data collection from a Student Academic Performance dataset, comprising both behavioral metrics and spatiotemporal information. The raw data undergoes preprocessing, including missing value imputation, one-hot encoding of categorical variables, and min-max scaling of numerical features. The processed data is then passed through two specialized branches: a Tabular Neural Structure-Aware (TabNSA) module that captures complex interdependencies within behavioral data, and a Spatiotemporal Transformer module that models temporal and sequential patterns in learning activities. The feature embeddings from both branches are fused and passed through fully connected layers to generate predictions across five academic performance bands, enabling precise classification and early risk identification. EduFuseNet achieved an accuracy of 99.00%, with a precision of 99.04%, recall of 99.00%, and F1-score of 99.01%, reflecting strong and reliable predictive performance. By leveraging both behavioral and temporal learning indicators, the model serves as an effective tool for early academic monitoring and intervention.

Abstract The number of manufacturing jobs in the US has been consistently increasing, driven by a rapidly evolving industrial landscape and the implementation of a new strategic plan. At the same time, concerns have emerged about the problem-solving abilities of engineering students, who represent the future workforce. This highlights the need for systematic evaluation and deeper insight into how these students approach problem-solving. In this paper, we introduce a virtual reality (VR)-based manufacturing environment combined with a data-driven analytical workflow to evaluate engineering students' problem-solving performance. Within the VR system, students complete assembly tasks to build car toys that meet specific design criteria. During the process, we capture real-time eye-tracking data, reflecting the spatial and temporal dynamics of their visual attention. We extract latent features from this data via a long short-term memory (LSTM)-based supervised representation learning for problem-solving performance evaluation. Our approach outperforms the traditional performance metrics-based evaluation by capturing the nonlinear dynamics from the in situ production process. Experimental results show that the learned feature representations provide significantly clearer distinctions in categorizing the students' problem-solving performance compared to the traditional performance metrics. The proposed evaluation framework holds broader potential for improving problem-solving assessments in various manufacturing systems and workforce training programs.

Behavioral, operational, and contextual factors affecting employee performance in the service industry are complex, and hence, predicting and intervening on them are especially difficult. As the banking industry is a crucial component of the service sector, it is particularly important to focus on predicting and intervening in the factors affecting the performance of banking employees. Current AI-based HR analytics solutions do not have a system for preserving fairness and optimizing decisions, as they tend to be based on a single source of signals. To solve these challenges, the present study introduces BEBOP-Net, a composite multi-source behavioral modeling and reinforcement learning agent that is used to elucidate expected employee behaviors and provide managerial control interventions that are comparatively inexpensive to implement. The structured HR features, chronological logs of the operational history, and written interaction records are modeled using a single multi-source behavior encoder with high temporal attention to extract dynamic patterns.

The circadian clock controls a vast array of cellular and organismal functions, from the molecular scale to behavior. While each cell is regimented by a cell-autonomous clock, few studies in the brain have dissected the circuit and behavioral contributions of cell-specific clocks. Relatedly, astrocytes are now known to play key roles in regulating synaptic function, circuit activity and behavior, but whether these functions are guided by astrocyte-autonomous clocks is unknown. Here, we report that post-natal deletion of the critical circadian clock gene Bmal1 in astrocytes, which abrogates core clock function in a cell type specific manner, induced expression of genes related to extracellular matrix (ECM) production, maintenance, and remodeling. Circadian variations have been shown in a specific ECM structure, perineuronal nets (PNNs), which are implicated in synaptic function and plasticity. In astrocyte-specific Bmal1 knockouts, hippocampal PNN abundance was decreased, and the circadian rhythm of these structures was also abolished. In line with evidence implicating PNNs, and the ECM in general, in synaptic function and plasticity, we found that astrocyte-specific Bmal1 KO mice had increased synaptic strength but blunted long term potentiation (LTP), as well as impaired learning and memory performance in a novel object recognition task. Taken together, these findings suggest that the astrocyte circadian clock regulates circadian rhythms in perineuronal net abundance as well as synaptic plasticity and behavioral learning and memory.

A quantitative approach with a causal design was used in this study to analyze the influence of Digital Behavior and Professional Learning Community on the Professionalism of Elementary School Teachers in Penajam Paser Utara Regency in the 2025/2026 academic year. The independent variables were defined as Digital Behavior (X1) and Professional Learning Community (X2), while the dependent variable was defined as Teacher Professionalism (Y). The study population consisted of 520 public elementary school teachers spread across four sub-districts. The sample was determined using the Slovin formula at a 7% error rate, resulting in 147 respondents and then rounded to 150 teachers to increase accuracy. The sampling technique was proportional random sampling. Data were collected using a five-level Likert-scale closed questionnaire. All instruments were tested for validity and reliability before use. Data analysis was performed using SPSS version 22 through classical assumption tests (normality and linearity) and multiple linear regression. Hypothesis testing was carried out at a significance level of 0.05 and strengthened by the coefficient of determination (R²). The results of the study show that partially and simultaneously, Digital Behavior and Professional Learning Community have a positive and significant influence on Teacher Professionalism. These findings imply that strengthening teachers’ digital competencies and fostering collaborative professional learning communities within schools are essential strategies to enhance teacher professionalism and improve the quality of elementary education.

Ergonomics or human factors engineering (HFE) mainly exploits human experiments to discover one's cognitive and behavioral mechanisms. Such a paradigm, however, suffers from the scale of subject group and the extent to which they can stand for the whole studied population. Additionally, for real-time human-machine tasks, the experiment-modeling-validation-application path may not be applicable since the experiment cannot be flexibly conducted to update cognitive models, leading to a failure of the online system control and management. To solve the dilemma, this article proposes the generative artificial intelligence (GAI)-driven ergonomics to augment the HFE research. By introducing GAI techniques, virtual-real hybrid experiments are combined and supplement more heterogeneous samples, enhancing the input diversity for cognitive modeling and behavioral learning. The case studies of human-machine cooperative driving and aerospace robotic arm operation indicate that the innovative paradigm can effectively and efficiently augment the human experiment data. It can elevate the generality and robustness of human models.

Cow-calf contact (CCC) systems are of growing scientific interest due to the public concern about early separation of the calf from the dam on dairy cattle farms, which currently occurrs soon after calving. Despite the scientifically demonstrated beneficial effects of CCC for both the cow (improving udder health and promoting maternal behaviors) and calf (growth, social learning and reduction of abnormal behaviors), only a small number of dairy cattle farmers practice CCC. Although unpredictable, new animal welfare legislation or best practice guidelines might suggest the adoption of CCC systems in the future. Therefore, the aim of this review is to identify common barriers that prevent farmers from implementing CCC systems and suggest suitable strategies to overcome them. Thus, this review focuses first on identifying barriers that prevent farmers from choosing CCC or transitioning to CCC, considering influencing factors such as region, demographics, and farm characteristics. Second, the review proposes suitable research and knowledge transfer strategies (i.e., communication, dissemination, and education) to address these barriers. For the proposal of research strategies, an overview of mother-offspring management strategies used in other livestock production systems (e.g., beef cattle and dairy sheep) that could potentially be used by dairy cattle farmers to implement CCC systems is provided. The results show that many dairy cattle farmers without CCC experience expressed concerns about the effects of CCC on economic viability, management and staff well-being, and animal welfare. However, farmers who practice(d) CCC reported generally positive experiences and did not confirm some of the concerns raised by farmers without CCC experience. This implies that these concerns are knowledge barriers that could be addressed with effective knowledge transfer strategies. Nevertheless, there were also concerns (e.g., separation distress) that were confirmed by farmers who practice(d) CCC, suggesting that these practical barriers require the exploration of novel strategies. Despite differences in biology or productive values, inspiration could be drawn from mother-offspring management in other livestock systems. For instance, future research on dairy cattle could explore the efficacy of restricted suckling to increase milk yield, as well as the efficacy of keeping calves together with familiar peers to reduce separation distress. Combined with research into economic viability and effective knowledge transfer, insights from other livestock systems could improve the implementation of CCC systems in a way that is sustainable for dairy cattle farmers, beneficial for the welfare of cows and calves, and socially acceptable.

This study systematically examines the interrelationship between green human resource management (GHRM), green innovation, and the blue economy in promoting sustainable coastal development in Indonesia. Using a Systematic Literature Review (SLR) approach, this research analyzes 65 peer-reviewed academic articles published between 2021 and 2025. The review identifies four major thematic clusters: (1) the implementation of GHRM practices, (2) the linkage between GHRM and green innovation, (3) the role of green innovation in socio-coastal contexts, and (4) the integration of GHRM and green innovation within the blue economy framework. The findings indicate that GHRM plays a critical role in shaping environmental competencies, pro-environmental behavior, and organizational learning, which collectively stimulate green innovation. Furthermore, the integration of GHRM and green innovation strengthens the socio-economic resilience of coastal communities while supporting sustainable marine resource management. However, the literature also reveals limited empirical application of GHRM in community-based coastal development, particularly within developing regions. This study contributes by offering a human-centered and knowledge-based conceptual framework for sustainable coastal development, providing strategic insights for policymakers, local governments, and development practitioners to enhance the implementation of the blue economy through strengthened human resource capacity and environmental innovation.

This study aims to analyze the impact of herding behaviour in mediating the relationship between social media exposure and financial literacy on the investment decisions of young investors affiliated with the UPN Veteran East Java Investment Gallery. This study involved 173 young investors selected using probability sampling. The research approach was quantitative, using an online questionnaire and analysis based on Partial Least Squares Structural Equation Modelling (SEM-PLS). The results of this study indicate a significant positive influence of social media exposure and financial literacy on investment decisions. Other findings indicate that herding behaviour is able to mediate complementarily in the influence of social media exposure on investment decisions and mediate competitively in the influence of financial literacy on investment decisions. This study is expected to enrich knowledge regarding the mediating role of herding behaviour within the framework of the Theory of Planned Behaviour and Social Learning Behaviour, which can emphasise the importance of financial literacy and digital literacy among young investors. The limitation of this study lies in the use of a limited sample at UPN Veteran East Jawa Investment Gallery, so it is recommended that future studies expand the scope of the research area.

The rapid digitization of education has transformed adult learning environments, with institutions increasingly deploying learning management systems, mobile learning platforms, and analytics-driven instructional tools to improve accessibility and flexibility. Despite these advancements, adult education programs continue to experience inconsistent technology adoption, uneven learner engagement, and significant dropout rates. Adult learners differ from traditional students in that they often balance education with employment, family responsibilities, and varying levels of digital literacy, which can influence their interaction with educational technologies. Consequently, understanding the determinants of technology adoption and sustained participation within adult education systems has become a critical research challenge. This study investigates technology adoption dynamics in adult education through predictive modeling of learner participation and retention patterns within digital learning environments. Using learner interaction logs, course engagement metrics, and demographic variables collected from adult education platforms, machine learning models are developed to predict technology usage intensity and the likelihood of learner persistence throughout course completion. The modeling framework evaluates behavioral indicators such as login frequency, content interaction, assignment submission patterns, and peer collaboration signals to identify predictors of retention outcomes. By linking behavioral learning analytics with predictive modeling, the study provides a systematic approach for identifying early risk signals of disengagement and technology abandonment. The findings offer actionable insights for designing adaptive adult education platforms that support sustained learner participation and improved retention outcomes.