Melanoma is a malignant cancer type which develops when DNA damage occurs (mainly due to environmental factors such as ultraviolet rays). Often, melanoma results in intense and aggressive cell growth that, if not caught in time, can bring one toward death. Thus, early identification at the initial stage is fundamental to stopping the spread of cancer. In this paper, a ViT-based architecture able to classify melanoma versus non-cancerous lesions is presented. The proposed predictive model is trained and tested on public skin cancer data from the ISIC challenge, and the obtained results are highly promising. Different classifier configurations are considered and analyzed in order to find the most discriminating one. The best one reached an accuracy of 0.948, sensitivity of 0.928, specificity of 0.967, and AUROC of 0.948.

Understanding how brain connectivity is disrupted across stages of traumatic brain injury (TBI) is essential for improving diagnosis and treatment. TBI poses major challenges in clinical assessment, requiring advanced neuroimaging and machine learning (ML) for effective patient stratification. This study classifies TBI patients into acute, chronic, and control groups using graph convolutional networks (GCNs) applied to structural connectomes derived from diffusion-weighted imaging (DWI). To enhance interpretability, Gradient-weighted Class Activation Mapping (Grad-CAM) was used to identify brain regions contributing to classification. The dataset included 40 participants: 18 acute TBI patients (Glasgow Coma Scale ≤ 6, enrolled after ≥ 24 hours of unresponsiveness), 6 chronic patients with persistent disorders of consciousness, and 16 healthy controls. Nine acute patients who regained consciousness were later included in the chronic group to assess longitudinal changes. The GCN was trained on DWI-derived connectomes and evaluated using leave-one-subject-out (subject-wise) cross-validation. It achieved 83.67% accuracy, with precision, recall, and F1-score of 81.6%, 78%, and 79%, respectively, reported as per-fold averages. Grad-CAM identified thalamus, anterior cingulate cortex, and frontal cortex as key regions for group discrimination. Results suggest a shift from widespread neural disruption in acute TBI to more localized impairments in the chronic stage, possibly reflecting compensatory reorganization. Despite the limited sample size, the model’s robustness was supported by conservative regularization and subject-wise validation. Notably, the GCN outperformed classical ML classifiers, offering superior accuracy and greater biological plausibility. These findings support the use of GCN-based pipelines for clinical decision support and highlight their potential to inform interpretable, ML-driven strategies for precision neurorehabilitation.

In this paper, we propose a Linux-based operating system, namely, DicomOS, tailored for medical imaging and enhanced interoperability, addressing user-friendly functionality and the main critical needs in radiology workflows. Traditional operating systems in clinical settings face limitations, such as fragmented software ecosystems and platform-specific restrictions, which disrupt collaborative workflows and hinder diagnostic efficiency. Built on Ubuntu 22.04 LTS, DicomOS integrates essential DICOM functionalities directly into the OS, providing a unified, cohesive platform for image visualization, annotation, and sharing. Methods include custom configurations and the development of graphical user interfaces (GUIs) and command-line tools, making them accessible to medical professionals and developers. Key applications such as ITK-SNAP and 3D Slicer are seamlessly integrated alongside specialized GUIs that enhance usability without requiring extensive technical expertise. As preliminary work, DicomOS demonstrates the potential to simplify medical imaging workflows, reduce cognitive load, and promote efficient data sharing across diverse clinical settings. However, further evaluations, including structured clinical tests and broader deployment with a distributable ISO image, must validate its effectiveness and scalability in real-world scenarios. The results indicate that DicomOS provides a versatile and adaptable solution, supporting radiologists in routine tasks while facilitating customization for advanced users. As an open-source platform, DicomOS has the potential to evolve alongside medical imaging needs, positioning it as a valuable resource for enhancing workflow integration and clinical collaboration.

The increasing prevalence of suicidal ideation among university students, especially those enrolled in medical and healthcare courses, necessitates innovative approaches for early detection and prevention. The southern part of Italy presents a distinct cultural, social, and economic context and is under‐represented in research. This study leverages the capabilities of machine learning methods and explainable artificial intelligence (XAI) to address this challenge, developing a predictive model for suicidal thoughts within a large sample of university students in Sicily, the southernmost and largest Italian region. The XGBoost algorithm was implemented for model training on survey data from 3278 students, including sociodemographic factors, academic performance metrics, and mental health indicators such as the Depression, Anxiety, and Stress Scale short version (DASS‐21) and the Eating Attitude Test short version (EAT‐26). The optimized XGBoost achieved promising performance, with AUROC ranging from 0.719 to 0.808 and balanced accuracy between 0.666 and 0.780 across internal and external test sets. Probability calibration was generally reasonable (Brier 0.151–0.185; ECE 0.035–0.120), supporting the reliability of predicted risk estimates. Performance in general varied by degree program, suggesting cohort‐specific characteristics. Our findings underscore the potential of machine learning models to identify at‐risk students and XAI for understating model prediction, emphasizing the importance of early intervention and targeted mental health support in university settings. The model′s performance and explainability highlight the potential in developing a student‐centric and self‐assessment framework while helping to mitigate potential ethical concerns associated with predictive mental health tools.

We introduce the Dendrite-Inspired Recurrent Unit (DIRU), a novel recurrent neural network architecture that incorporates compartmentalized computation inspired by the hierarchical processing of biological dendrites. Unlike classical recurrent models, which encode temporal dynamics within a single homogeneous hidden state, DIRU implements multiple interacting nonlinear compartments within each recurrent unit, enabling structured internal computation, hierarchical state formation, and adaptive dynamical processing. We evaluate DIRU on canonical chaotic systems, including the Lorenz attractor and the Mackey–Glass time-delay system, using multi-step forecasting tasks that explicitly probe long-horizon stability, error growth consistent with bounded chaotic dynamics, and regime preservation. Across all benchmarks, DIRU demonstrates consistently lower prediction error and more stable training dynamics than baseline models across the evaluated settings, with statistically supported performance differences under controlled capacity conditions. To assess real-world biological relevance, we further evaluate DIRU on neonatal electroencephalography (EEG) seizure classification using the Helsinki Neonatal EEG dataset. Operating directly on raw EEG time-series windows under extreme class imbalance, DIRU achieves improved discriminative performance relative to baseline architectures (AUC = 0.884), demonstrating robust detection of transient pathological dynamics without aggressive feature engineering or data augmentation. Together, these results establish multi-compartment recurrent computation as a powerful inductive bias for learning complex nonlinear systems. DIRU provides a promising framework for modeling chaotic dynamics and biological time series, bridging neuroscience-inspired computation with modern deep sequence learning, and offering a principled pathway toward structured recurrent architectures for complex dynamical data.