Objective: Student mental health has emerged as an increasingly prominent issue in sustainable educational healthcare systems. Accurately and promptly identifying students' depression and analyzing the key factors associated with it are crucial for improving student mental health. Method: We propose an artificial intelligence algorithm, GLNet, that integrates Mamba and convolutional layers to extract features from students' demographic, academic, and lifestyle information for depression analysis. The performance of GLNet is validated on the publicly available Student Depression Dataset. Results: GLNet achieves an accuracy of 88.84% on the Student Depression Dataset, outperforming other methods and verifying its effectiveness in student depression analysis. Factor contribution analysis indicates that academic pressure and financial stress may be associated with student depression, while healthy dietary habits and academic satisfaction may alleviate depression. Subgroup analysis further reveals that a higher Cumulative Grade Point Average may be positively correlated with depression in female students, and unhealthy dietary habits may be linked to depression among doctoral students. Conclusion: GLNet can serve as a reliable tool for enhancing student mental health. It also provides valuable insights for educators to identify students at risk of depression, contributing to the optimization of student mental health intervention strategies.
Large-scale 3D mapping and high-resolution remote sensing are essential for environmental monitoring, disaster assessment, and urban planning. Heterogeneous unmanned aerial vehicle (UAV) swarms, equipped with complementary sensing and onboard edge computing capabilities, offer efficient, adaptive, and resource-aware operations. However, achieving complete spatial coverage, ensuring sensing relevance, and optimizing both communication and computational resources remain challenging under dynamic and complex conditions. This paper proposes an energy- and resource-aware cooperative framework, DMMP-PR-TSA, which integrates remote sensing data-driven region partitioning, improved self-organizing map (SOM)-based intelligent pre-assignment, priority-aware dynamic task reallocation (PR), and reinforcement learning (RL)-based task sequence adjustment (TSA). The framework jointly optimizes spatial path planning for sensing tasks and computational resource allocation for edge processing and collaborative task execution, while embedding priority handling to meet deadlines for critical missions. Compared with baseline algorithms, DMMP-PR-TSA demonstrates \(15\%\!-\!20\%\) higher completion rates in large-scale missions, \(10\%\!-\!30\%\) improvement under dynamic fleet changes, and consistently higher success rates for high-priority tasks. Simulation results validate its scalability, robustness, and mission-critical applicability, highlighting its effectiveness in advancing the intelligence and operational efficiency of UAV-based large-scale remote sensing and edge-computing-assisted systems.
12 January 2026 in Frontiers in Behavioral Neuroscience
Objective: Gliomas, the most aggressive type of brain tumor, are infamous for their low survival rates. Tumor grading and isocitrate dehydrogenase (IDH) status are key prognostic biomarkers for gliomas. However, obtaining these markers typically requires invasive methods such as biopsy. As an effective, noninvasive alternative, multimodal MRI can reveal tumor spatial information and the microenvironment. Low-grade and IDH-mutant gliomas often exhibit T2-FLAIR mismatch signals. Medical image foundational models can explore complex representations in medical images, and fine-tuning them may further enhance glioma diagnosis. Methods: We propose a multi-task network, MTSAM, for simultaneous glioma IDH genotyping and grading. MTSAM first uses dilated convolutions to simulate large-field convolutions and then reviews the T2 and FLAIR images. Then, we employ convolutions to perform a detailed exploration of the T2 and FLAIR images, and we subtract the weighted T2 and FLAIR images to obtain T2-FLAIR mismatch features. T2-FLAIR mismatch features are concatenated with multimodal MRIs and input into the customized SAM-Med3D. The customized SAM-Med3D is fine-tuned by leveraging complementary information across multi-view modalities, including MRIs, handcrafted radiomics (HCR), and clinical features. Then it extracts deep features for accurate IDH genotyping and grading. Results: MTSAM achieves AUCs of 92.38 and 94.31% for glioma IDH typing and grading on the UCSF-PDGM dataset, respectively, and AUCs of 91.56 and 93.37% on the BraTS2020 dataset, outperforming other methods. Additionally, we use Grad-CAM to visualize the attention maps of MTSAM, demonstrating its potential for non-invasive glioma diagnosis. Conclusion: The proposed method demonstrates that we can effectively fuse multi-view, non-invasive information and fully explore the knowledge learned by medical image foundational models from large-scale medical datasets to facilitate glioma diagnosis, thereby advancing glioma research.
Introduction: With the widespread application of artificial intelligence (AI) technology in design education, exploring how AI tools shape students’ innovative abilities has become increasingly important. Existing technology acceptance models mainly explain the adoption behavior of AI tools but have not examined how technological features influence innovation outcomes through user psychological processes. Methods: This study employs a cross-sectional quantitative design to examine how AI tool quality dimensions influence design innovation ability (DIA). Specifically, it tests the pathways through which interaction quality (IQT) and information quality (INQ) affect DIA via satisfaction (SAT) and intention to use (INU). Based on an online survey of 1,016 Chinese industrial design students, PLS-SEM data analysis was employed. Results: The study found that both IQT and INQ positively influence SAT. Subsequently, SAT, IQT, and INQ jointly affect INU. INU significantly predicts DIA. Mediating effect analysis confirmed that SAT plays a partial mediating role between the quality dimensions and INU. Discussion: Notably, the study reveals a pattern in which information quality exerts stronger effects than interaction quality, and direct functional evaluation outweighs affective mediation, both of which challenge conventional technology acceptance assumptions. The findings extend technology acceptance theory by identifying dual mechanisms, including direct functional and indirect emotional pathways, in AI-assisted creative education.
With the growing use of in-body health sensors (IBHS) for real-time care, large-scale data processing, mobile edge computing (MEC) support, and sustainable power have become key bottlenecks. IBHS are long-term implants, so conventional recharging is infeasible; radio-frequency wireless energy transfer (RF-WET) is the primary means of continuous power. During data uploading and energy harvesting, adversaries can passively observe application-layer signals such as offloading behavior, charging frequency, and energy states to infer user activities and locations, creating serious privacy risks. We propose a deep reinforcement learning (DRL) based task offloading framework that jointly considers MEC, energy awareness, and privacy preservation, and we introduce an energy-aware privacy entropy model to quantify risk across the offloading process. By jointly sensing wireless channels, MEC resource availability, and energy status, the scheme dynamically optimizes data uploading and computation offloading, while analyzing location-privacy leakage induced by RF-WET access points (WET-APs). Simulations show clear gains over state-of-the-art baselines in latency, energy efficiency, and location-privacy protection, indicating strong potential for next-generation edge-assisted healthcare systems.
Background: Childhood maltreatment is a well-established risk factor for various mental health and behavioral problems, including non-suicidal self-injury (NSSI). Despite its prevalence, the associations between childhood maltreatment and NSSI, particularly as mediated by alexithymia and depression, remain not fully explored. This study primarily employs structural equation modeling (SEM) to test the hypothesized relationships between childhood maltreatment, alexithymia, depression, and NSSI, and additionally incorporates network analysis to examine the symptom-level connections among these constructs at the micro level, complementing the macro-level insights from SEM. Method: We recruit a sample of 1,414 college students. Participants complete the Childhood Trauma Questionnaire-Short Form (CTQ-SF), the Toronto Alexithymia Scale (TAS-20), the Self-Rating Depression Scale (SDS), and the NSSI Behavior Assessment Questionnaire. We then employ SEM to test the hypothesized mediation model and use network analysis to visualize the interconnections among the specific dimensions of childhood maltreatment, alexithymia, depression, and NSSI. Results: SEM results indicate that alexithymia and depression significantly mediate the relationship between childhood maltreatment and NSSI, forming a chain mediating effect. Network analysis identifies “Difficulty identifying feelings” (DIF) as the most central node in the network. Moreover, DIF, NSSI, and Affective Symptoms (AS) were identified as key bridge symptoms connecting the clusters within the network. Conclusion: Our findings highlight the importance of focusing on the emotional processing deficit and emotional problems in the intervention and prevention of NSSI among college students.