Over the past decade, deep learning has achieved unprecedented successes in a diversity of application domains, given large-scale datasets. However, particular domains, such as healthcare, inherently suffer from data paucity and imbalance. Moreover, datasets could be largely inaccessible due to privacy concerns, or lack of data-sharing incentives. Such challenges have attached significance to the application of generative modeling and data augmentation in that domain. In this context, this study explores a machine learning-based approach for generating synthetic eye-tracking data. We explore a novel application of variational autoencoders (VAEs) in this regard. More specifically, a VAE model is trained to generate an image-based representation of the eye-tracking output, so-called scanpaths. Overall, our results validate that the VAE model could generate a plausible output from a limited dataset. Finally, it is empirically demonstrated that such approach could be employed as a mechanism for data augmentation to improve the performance in classification tasks.

Background: During the coronavirus disease 2019 (COVID-19) pandemic, calculation of the number of emergency department (ED) beds required for patients with vs. without suspected COVID-19 represented a real public health problem. In France, Amiens Picardy University Hospital (APUH) developed an Artificial Intelligence (AI) project called “Prediction of the Patient Pathway in the Emergency Department” (3P-U) to predict patient outcomes. Materials: Using the 3P-U model, we performed a prospective, single-center study of patients attending APUH’s ED in 2020 and 2021. The objective was to determine the minimum and maximum numbers of beds required in real-time, according to the 3P-U model. Results A total of 105,457 patients were included. The area under the receiver operating characteristic curve (AUROC) for the 3P-U was 0.82 for all of the patients and 0.90 for the unambiguous cases. Specifically, 38,353 (36.4%) patients were flagged as “likely to be discharged”, 18,815 (17.8%) were flagged as “likely to be admitted”, and 48,297 (45.8%) patients could not be flagged. Based on the predicted minimum number of beds (for unambiguous cases only) and the maximum number of beds (all patients), the hospital management coordinated the conversion of wards into dedicated COVID-19 units. Discussion and conclusions: The 3P-U model’s AUROC is in the middle of range reported in the literature for similar classifiers. By considering the range of required bed numbers, the waste of resources (e.g., time and beds) could be reduced. The study concludes that the application of AI could help considerably improve the management of hospital resources during global pandemics, such as COVID-19.

The availability of data is a key enabler for researchers across different disciplines. However, domains, such as healthcare, are still fundamentally challenged by the paucity and imbalance of datasets. Health data could be inaccessible due to a variety of hurdles such as privacy concerns, or lack of sharing incentives. In this regard, this study aims to publish an eye-tracking dataset developed for the purpose of autism diagnosis. Eye-tracking methods are used intensively in that context, whereas abnormalities of the eye gaze are largely recognised as the hallmark of autism. As such, it is believed that the dataset can allow for developing useful applications or discovering interesting insights. As well, Machine Learning is a potential application for developing diagnostic models that can help detect autism at an early stage of development.

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition characterised by challenges in social communication, repetitive behaviours, and restricted interests [1]. Early and accurate diagnosis is critical for effective intervention, enabling individuals with ASD to achieve better developmental outcomes and an improved quality of life. However, traditional diagnostic methods, often reliant on subjective behavioural observations, remain timeintensive and inconsistently accessible. This underscores an urgent need for innovative, scalable, and objective diagnostic tools [2,3].Machine Learning (ML) has emerged as a transformative approach to ASD diagnosis, offering the ability to analyse large, complex datasets and uncover patterns that surpass human capability. For instance, eye-tracking technologies have been extensively utilised to quantify gaze behaviours such as fixations and saccades, as well-established markers of autism. Studies employing Deep Learning have achieved high accuracy in classifying ASD from typically developing individuals based on eye-tracking data [3,7]. These technological advancements provide a foundation for developing tools that are not only efficient but also potentially generalisable across diverse populations.Furthermore, approaches such as transforming gaze scanpaths into visual representations for classification have simplified the diagnostic pipeline, enabling the automation of traditionally laborious processes [4]. Additionally, unsupervised learning techniques, including clustering of eye-tracking data, have demonstrated potential for extracting unique insights into the variability of ASD presentations [5]. These developments illustrate the growing synergy between AI-driven tools and clinical practices.Beyond eye tracking, other data modalities such as structural MRI (sMRI), resting-state functional connectivity (rsFC), and multimodal approaches integrating genetic, behavioural, and imaging data have shown promise in identifying robust biomarkers for ASD. These diverse methodologies underscore the importance of leveraging multidimensional data to improve diagnostic precision and reliability [2,6]. Despite these promising innovations, challenges persist. Standardisation of methodologies, reproducibility of results, and the translation of research into clinical applicability remain significant barriers. This special issue seeks to address these challenges by presenting cuttingedge research that integrates ML and neuroinformatics to enhance the accuracy, efficiency, and accessibility of ASD diagnostics. By bridging the gap between technology and practice, this collection of studies aims to drive the field toward more effective and equitable solutions for ASD diagnosis.The articles included in this issue explore various aspects of ASD diagnosis through ML, presenting innovative approaches and significant findings:Eslami et al. provide a comprehensive review of ML models applied to sMRI and fMRI datasets, examining their efficacy in diagnosing ASD and related disorders. The study highlights key advancements in deep learning architectures and identifies limitations such as data heterogeneity and reproducibility challenges. https://www.frontiersin.org/journals/neuroinformatics/articles/10.3389/fninf.2020.575999Vector Machine (SVM) models. Their research uncovers discriminative connectivity patterns within the Default Mode Network (DMN), achieving high classification accuracy and reinforcing the potential of rsFC data in ASD diagnostics. https://www.frontiersin.org/journals/neuroinformatics/articles/10.3389/fninf.2022.761942Jia et al. conduct a bibliometric analysis, mapping the global research landscape of AI applications in ASD. Their findings highlight trends such as the rise of feature selection and the significance of multimodal integration, providing a roadmap for future studies. https://www.frontiersin.org/journals/neuroinformatics/articles/10.3389/fninf.2023.1310400Ruan et al. present an exploratory study on using micro-expressions as diagnostic biomarkers. Despite the challenges posed by video data quality, their work emphasises the need for multimodal approaches combining behavioural and neuroimaging data. https://www.frontiersin.org/journals/neuroinformatics/articles/10.3389/fninf.2024.1435091The contributions in this issue emphasise the multidimensional nature of ASD and the need for a holistic diagnostic framework. Challenges such as the lack of standardisation across datasets, ethical considerations in algorithm deployment, and the interpretability of ML models remain relevant. However, the integration of advanced computational methods with clinical expertise opens avenues for personalised treatment strategies and early intervention protocols.We envision that future research should focus on:• Data Diversity and Multimodal Integration: Combining imaging, genetic, and behavioural data to enhance model robustness.• Interpretable AI: Developing transparent algorithms that clinicians can trust and use effectively.

Background: Overcrowded emergency departments (EDs) create significant challenges for patient management and hospital efficiency. In response, Amiens Picardy University Hospital (APUH) developed the “Prediction of the Patient Pathway in the Emergency Department” (3P-U) model to enhance patient flow management. Objectives: To develop and clinically validate an explainable artificial intelligence (XAI) model for hospital admission predictions, using structured triage data, and demonstrate its real-world applicability in the ED setting. Methods: Our retrospective, single-center study involved 351,019 patients consulting in APUH’s EDs between 2015 and 2018. Various models (including a cross-validation artificial neural network (ANN), a k-nearest neighbors (KNN) model, a logistic regression (LR) model, and a random forest (RF) model) were trained and assessed for performance with regard to the area under the receiver operating characteristic curve (AUROC). The best model was validated internally with a test set, and the F1 score was used to determine the best threshold for recall, precision, and accuracy. XAI techniques, such as Shapley additive explanations (SHAP) and partial dependence plots (PDP) were employed, and the clinical explanations were evaluated by emergency physicians. Results: The ANN gave the best performance during the training stage, with an AUROC of 83.1% (SD: 0.2%) for the test set; it surpassed the RF (AUROC: 71.6%, SD: 0.1%), KNN (AUROC: 67.2%, SD: 0.2%), and LR (AUROC: 71.5%, SD: 0.2%) models. In an internal validation, the ANN’s AUROC was 83.2%. The best F1 score (0.67) determined that 0.35 was the optimal threshold; the corresponding recall, precision, and accuracy were 75.7%, 59.7%, and 75.3%, respectively. The SHAP and PDP XAI techniques (as assessed by emergency physicians) highlighted patient age, heart rate, and presentation with multiple injuries as the features that most specifically influenced the admission from the ED to a hospital ward. These insights are being used in bed allocation and patient prioritization, directly improving ED operations. Conclusions: The 3P-U model demonstrates practical utility by reducing ED crowding and enhancing decision-making processes at APUH. Its transparency and physician validation foster trust, facilitating its adoption in clinical practice and offering a replicable framework for other hospitals to optimize patient flow.

Background/Objectives: Endometriosis is a chronic inflammatory condition that often requires laparoscopic examination for definitive diagnosis. Automated analysis of laparoscopic images using Deep Learning (DL) may support clinicians by improving diagnostic consistency and efficiency. This study aimed to develop and evaluate explainable DL models for the binary classification of endometriosis using laparoscopic images from the publicly available GLENDA (Gynecologic Laparoscopic ENdometriosis DAtaset). Methods: Four representative architectures—ResNet50, EfficientNet-B2, EdgeNeXt_Small, and Vision Transformer (ViT-Small/16)—were systematically compared under class-imbalanced conditions using five-fold cross-validation. To enhance interpretability, Gradient-weighted Class Activation Mapping (Grad-CAM) and SHapley Additive exPlanations (SHAP) were applied for visual explanation, and their quantitative alignment with expert-annotated lesion masks was assessed using Intersection over Union (IoU), Dice coefficient, and Recall. Results: Among the evaluated models, EdgeNeXt_Small achieved the best trade-off between classification performance and computational efficiency. Grad-CAM produced spatially coherent visualizations that corresponded well with clinically relevant lesion regions. Conclusions: The study shows that lightweight convolutional neural network (CNN)–Transformer architectures, combined with quantitative explainability assessment, can identify endometriosis in laparoscopic images with reasonable accuracy and interpretability. These findings indicate that explainable AI methods may help improve diagnostic consistency by offering transparent visual cues that align with clinically relevant regions. Further validation in broader clinical settings is warranted to confirm their practical utility.