Background: Deep learning has demonstrated significant potential in advancing computer-aided diagnosis for neuropsychiatric disorders, such as migraine, enabling patient-specific diagnosis at an individual level. However, despite the superior accuracy of deep learning models, the interpretability of image classification models remains limited. Their black-box nature continues to pose a major obstacle in clinical applications, hindering biomarker discovery and personalized treatment. Objective: This study aims to investigate explainable artificial intelligence (XAI) techniques combined with multiple functional magnetic resonance imaging (fMRI) indicators to (1) compare their efficacy in migraine classification, (2) identify optimal model-indicator pairings, and (3) evaluate XAI’s potential in clinical diagnostics by localizing discriminative brain regions. Methods: We analyzed resting-state fMRI data from 64 participants, including 21 (33%) patients with migraine without aura, 15 (23%) patients with migraine with aura, and 28 (44%) healthy controls. Three fMRI metrics—amplitude of low-frequency fluctuation, regional homogeneity, and regional functional connectivity strength (RFCS)—were extracted and classified using GoogleNet, ResNet18, and Vision Transformer. For comprehensive model comparison, conventional machine learning methods, including support vector machine and random forest, were also used as benchmarks. Model performance was evaluated through accuracy and area under the curve metrics, while activation heat maps were generated via gradient-weighted class activation mapping for convolutional neural networks and self-attention mechanisms for Vision Transformer. Results: The GoogleNet model combined with RFCS indicators achieved the best classification performance, with an accuracy of >98.44% and an area under the receiver operating characteristic curve of 0.99 for the test set. In addition, among the 3 indicators, the RFCS indicator improved accuracy by approximately 8% compared with the amplitude of low-frequency fluctuation. Brain activation heat maps generated by XAI technology revealed that the precuneus and cuneus were the most discriminative brain regions, with slight activation also observed in the frontal gyrus. Conclusions: The use of XAI technology combined with brain region features provides visual explanations for the progression of migraine in patients. Understanding the decision-making process of the network has significant potential for clinical diagnosis of migraines, offering promising applications in enhancing diagnostic accuracy and aiding in the development of new diagnostic techniques.

To address the existing problems in emotion recognition research based on EEG data and improve recognition accuracy, so as to provide a better approach for scientific research and mental health maintenance.This paper presents an emotion recognition method tailored for small sample training data, effectively addressing the issue of emotion classification based on multichannel continuous electroencephalogram (EEG) signals. The proposed method first extracts differential entropy features from EEG signals and maps them into two-dimensional matrices based on their spectral characteristics. These pre-processed original data and five different frequency band matrices are then stacked into three-dimensional multichannel data, which are fed into the 3D-EEGU-Net model for classification. Experiments conducted on the public SEED dataset demonstrate an emotion recognition accuracy of 92.34%. Furthermore, to enhance the accuracy of emotion recognition for specific populations, the size of data blocks (patches) and model hyperparameters were dynamically adjusted. In the valence-based emotion ternary classification task conducted on the DEAP public dataset, the method achieved an emotion recognition accuracy of 72.02%. Comparative experiments show that the proposed method exhibits significant superiority in emotion recognition performance on both datasets. This research not only provides a new approach to EEG emotion recognition but also holds substantial potential for applications in detecting psychological disorders and physical conditions in individuals from specialized occupational groups.

Infectious diseases can sometimes lead to pandemics, often transmitted through public and social gatherings, including in-person hospital visits. Consequently, there is an urgent need for innovative approaches to prevent their spread. Taking COVID-19 as an example, we have explored a remote, contactless hospital online model that offers the public online medical consultations, professional psychological counseling, and chronic disease management consultations, thereby mitigating the risk of new transmissions resulting from hospital visits. This model was implemented, validated, and practiced at West China Hospital in China from 29 January 2020, to 12 March 2020. It was also applicable to other infectious diseases, such as influenza A. In this research, we utilized the hospital’s internet platform, supplemented by telephone services, to offer the following to the public: 1) General medical education and consultation related to epidemics and psychological anxiety; 2) Online screening for at-risk populations; 3) Online prescription and medication delivery services for patients with chronic diseases. Consequently, over a period of more than 1 month, the online epidemic platform completed a total of 32,755 cases, including 8,783 internet consultations and 1,082 telephone consultations for the public, as well as 22,890 internet consultations for chronic disease patients. Among these, 289 high-risk individuals were identified, with 3 cases confirmed as COVID-19 during follow-up diagnoses, while no infections were detected in the remaining individuals. In conclusion, this innovative medical model serves as a significant supplement to existing healthcare systems and has the potential to be expanded to other hospitals and other infectious diseases. It is particularly beneficial in scenarios where medical resources are limited, populations are under quarantine, and there is a large demand for medical services and anxiety management during infectious disease pandemics.

As a highly dense area for passengers, the airport departure hall is particularly characterized by the easy gathering of passengers, irregularity, and difficulty in command, which has become a high incidence of emergencies. For the evacuation problem in case of emergencies, the vast majority of passengers randomly choose the shortest path from the exit to escape, which will lead to congestion at some nodes and prolonging of the evacuation time. Therefore, this study proposes a better evacuation method considering the dynamic change in the spatiotemporal congestion of passengers. First, this paper constructs an evacuation network diagram for the departure hall of the Shenzhen Bao’an International Airport terminal building, simplifies the evacuation process into point-to-point evacuation, and proposes a time-varying network algorithm to optimize the emergency evacuation of the airport. Second, on the premise of obtaining the number and dynamic distribution of passengers in each area in real time, a linear programming model with the shortest evacuation time as the optimization objective is established, and the current optimal evacuation path of passengers is dynamically calculated. Finally, the two evacuation methods, random evacuation and evacuation, based on a time-varying network algorithm are simulated and analyzed by AnyLogic software. The results show that the evacuation scheme based on the time-varying network algorithm can effectively relieve the passenger congestion of some evacuation paths and safety exits. The evacuation time is reduced by 74 s, the evacuation efficiency is improved by 14%, passenger congestion at some evacuation routes and safety exits can be effectively relieved, and the global optimization of the emergency evacuation of passengers is achieved.