MethodsX最新文献

Depression affects over 280 million people worldwide, with neurological patients particularly prone to medication-induced episodes. Conventional diagnostic approaches rely on subjective evaluations, limiting reproducibility and consistency in clinical settings. This study proposes an interpretable deep learning framework for objective depression detection using EEG signals. We hypothesize that combining EEG-based features with explainable artificial intelligence can provide both high accuracy and transparency in diagnosis. The model was trained on EEG data from 232 neurological patients, achieving 98 % classification accuracy. Interpretability was enhanced through SHAP (SHapley Additive exPlanations) analysis, which identified clinically meaningful EEG biomarkers such as the delta/alpha ratio and theta band power. This paper highlights the following contributions: Integration of EEG features with a lightweight deep learning model for depression detection High diagnostic accuracy achieved while maintaining interpretability for clinicians An objective tool that is compatible with existing EEG infrastructure, supporting clinical adoption These results show that our framework bridges predictive performance with interpretability, offering a transparent and scalable EEG-based diagnostic tool. We conclude that this approach can complement clinical decision-making, reducing dependence on subjective evaluation and enabling more consistent, data-driven mental health care.

Image Captioning is a crucial task that enables systems to generate descriptive sentences for visual content. Though image captioning systems bloom at the intersection of Computer Vision and Natural Language Processing, these models act mostly as black boxes offering little or no insight into how captions are derived. We present a novel explainable image captioning framework that integrates a Convolutional Neural Network encoder with a Transformer decoder. Attention-based heatmaps are used to explain the visuals offering transparency in the decision making process. The method evaluates captioning quality and interpretability on the MS COCO dataset using BLEU, METEOR, CIDER and SPICE. The method enhances the trustworthiness and transparency, making it reliable for applications like healthcare, education, security, surveillance and forecasting. A reproducible method for integrating visual explainability into image captioning exploring transformer decoder attention maps. The method contributes to the growing body of eXplainable AI (XAI) by addressing the transparency gap in vision-language models Balance performance with interpretability paving the way for more transparent and trustworthy AI systems.

Antiepileptic drugs are considered contaminants of emerging concern in water and are resistant to conventional wastewater treatment processes. Therefore, their presence has been detected in surface waters, and their elimination/degradation requires effective treatment methods. In this research, ultrasound-based methods (e.g., sonolysis, sono-Fenton, and sono-photo-Fenton) were addressed in the degradation of antiepileptic drug primidone at laboratory scale. A high-frequency ultrasound (at 578 kHz and 20.4 W) was applied. Then, Fe²⁺ ions (5 mg l^-1) and a UVA lamp (4 W) were added to the sonochemical reactor. After 75 min of treatment, the sono-photo-Fenton method showed better degradation efficiency (93 %) than the sono-Fenton (83 %) and sonolysis (62 %) methods. Finally, the effectiveness of the degradation method by sono-photo-Fenton was tested in simulated pharmaceutical wastewater, degrading 72 % of primidone at 75 min of treatment, indicating matrix effect plays a role in the degradation (which could be a potential application of ultrasound hybridized with the photo-Fenton process).•Three ultrasound-based treatment methods were applied to degrade primidone in water.•The sono-photo-Fenton method degraded 93 % of primidone during 75 min of treatment.•The matrix influence on primidone degradation by sono-photo-Fenton was evaluated.

This article presents the development of LOSARI, a novel R-based statistical software designed to facilitate students' self-regulated learning (SRL) in statistics courses. LOSARI can be accessed online without installation and allows students to perform statistical analyses through a point-and-click interface without coding. It integrates several innovative features: interactive video tutorials embedded in the analysis environment, real-time error notifications that guide students in correcting mistakes, and automatic interpretation of results to support independent learning. The software was validated through a student satisfaction survey using the End-User Computing Satisfaction (EUCS) model, which indicated that most users had positive perceptions of LOSARI and found it effective for learning statistics outside the classroom. Possible extensions and enhancements are also discussed.•A structured process for developing LOSARI as an R-based statistical learning tool.•Introduction of key features, including interactive video tutorials, real-time error notifications, and automatic interpretation.•Validation method through student satisfaction measurement and comparison with manual statistical coding.

Ecosystem condition can be understood as the quality of an ecosystem in terms of its abiotic, biotic, and landscape characteristics. It is a measure of structural integrity, functional capacity, and resilience of any given ecological system. Its assessment is essential to support environmental objectives (e.g., nature restoration or sustainable use). Spatially explicit assessment of ecosystem condition requires integrating diverse geospatial data. Here, we present the EcoCondition Toolset, a QGIS plugin implementing a user-friendly GIS weighted-sum methodology for ecosystem condition assessments. It simplifies data preparation and analysis through five sequential toolsets: i) layer alignment and resampling; ii) no-data handling; iii) multicollinearity testing; iv) indicator normalisation and inversion; and v) condition assessment. The plugin calculates six specific ecosystem attribute - or state - composites (Physical, Chemical, Compositional, Structural, Functional, Landscape) from user-selected variables (in raster format), according to the System of Environmental-Economic Accounting. After data preparation and verification, the tool displays default equal weights for each composite and related variables, which users can adjust (e.g., to reflect stakeholder preferences). The toolset automates best-practice multicollinearity screening, normalisation, and flexible weighting for ecosystem condition assessment and monitoring. The resulting index preserves true severity and variation among ecosystem states. The results can support robust policy instruments and land-use decision-making, prioritising conservation and restoration actions.

Hand injuries are among the most common musculoskeletal injuries and can significantly impair an individual's ability to perform activities of daily living (ADL), thereby impacting quality of life. Self-efficacy plays a vital role in influencing daily performance and recovery following injury. This cross-sectional study aims to explore the relationship between ADL performance and self-efficacy among clients with hand injuries within the Indian context. Secondary objectives of this study include assessing self-efficacy levels and evaluating ADL performance in this population. • A self-administered, closed-ended, structured questionnaire comprising performance-based and self-efficacy measures will be used for data collection. Participants will include adults aged 18 years and above who have sustained fractures of the hand or wrist, including digits, and have undergone surgical treatment. • Clients will be recruited from the Occupational Therapy department. • The findings aim to highlight the importance of considering both objective and subjective measures in occupational therapy assessment and to emphasize the role of self-efficacy in ADL performance following hand injuries, potentially informing culturally sensitive rehabilitation interventions.

This study proposes an advanced spatio-temporal framework to forecast strategic food commodity prices in Indonesia using Geographically and Temporally Weighted Spline Regression (GTWSR), a nonparametric extension of GTWR designed to capture nonlinear spatio temporal effects. Monthly data from the Strategic Food Price Information Center (SFPIC) and Statistics Indonesia (BPS), covering eight key commodities and the Farmer Price Index across 34 provinces (January 2022-August 2024), were analyzed through spatial distance measurement, bandwidth optimization, local parameter estimation, and statistical validation. The GTWSR model demonstrated strong predictive performance (overall accuracy: R² = 91.61 %, RMSE = 1.22, MAE = 0.94, MAPE = 3.7 %), with rice and garlic achieving the highest accuracy, while red and cayenne chili showed greater errors due to price volatility. Spatial disparities were evident, as eastern provinces such as Papua, Maluku, and East Nusa Tenggara consistently faced higher prices compared to western regions. These findings underscore the need for region-specific interventions to strengthen logistics and stabilize horticultural supply chains. Limitations include reliance on monthly aggregated data, the temporal scope ending in 2024, and dependence on secondary datasets, which may affect replication and long-term applicability.