Zhangzhou Normal University

Objective: Cross-subject generalization in EEG-based seizure detection remains a significant challenge due to diverse physiological differences among patients, which can degrade model performance across individuals. This work aims to develop a robust, lightweight, and fast-deployable model for cross-subject seizure detection that achieves high accuracy while enabling real-time, low-latency inference on resource-constrained devices. Methods: We propose PSD-LW-DCN, a novel lightweight deep convolutional network that leverages power spectral density (PSD) features extracted via multitaper spectral estimation. Raw EEG signals are preprocessed through bandpass filtering, segmentation, and class balancing. PSD features are computed across multiple frequency bands ($$\delta$$,$$\theta$$,$$\alpha$$,$$\beta$$,$$\gamma$$), averaged across channels to enhance generalizability, and concatenated as input to the network. The model was trained and evaluated separately on the CHB-MIT and TUSZ datasets, with no data mixing between them. Results: With only 61,218 parameters, PSD-LW-DCN achieves state-of-the-art efficiency, reducing inference time to 1.9 ms/sample on CHB-MIT and 2.1 ms/sample on TUSZ, demonstrating its suitability for real-time applications. It attains 85.84% accuracy on CHB-MIT and 83.21% on TUSZ, representing improvements of 5% and 3%, respectively, over previous approaches. Notably, the model shows strong cross-subject generalization and robustness across diverse patient populations. In clinically meaningful event-based evaluation, it achieves a low false alarm rate of 0.33 false alarms per hour (FA/h) on CHB-MIT and 1.09 FA/h on TUSZ, indicating high operational reliability in long-term monitoring scenarios. Conclusions: PSD-LW-DCN offers an optimal balance between accuracy, speed, and model size, making it highly suitable for real-time, edge-based clinical diagnosis of epilepsy. Its efficiency and low false alarm burden confirm its potential for deployment in low-power wearable and point-of-care systems. Future work will focus on further reducing the false alarm rate, particularly in more heterogeneous datasets like TUSZ, to enhance clinical usability.

This paper investigates the problem of distributed resource management in underwater acoustic communication networks (UACNs) involving multiple transmitters and receivers. In this setting, each transmitter autonomously selects a power allocation strategy based solely on local observations, without reliance on a central controller. Given that the optimization problem incorporating fairness and quality of service (QoS) constraints is non-convex and NP-hard, it is reformulated as a Markov Decision Process (MDP). To address the high complexity of underwater networks and the large state and action spaces, we propose a distributed learning framework based on a multi-agent dueling deep Q-network (MAD3QN). The proposed scheme enables each transmitter to dynamically adjust its transmission power based on local observations by integrating the Jain fairness index, QoS interruption penalty, and energy consumption constraints. Furthermore, by incorporating a dueling network architecture and a neighborhood cooperation mechanism, the learning efficiency is significantly enhanced, leading to a stable and effective resource optimization policy. Simulation results demonstrate that the proposed distributed learning algorithm outperforms existing approaches in terms of convergence speed, network fairness, and communication rate.

Job crafting has become an essential strategy for kindergarten teachers to cope with increasing job demands and sustain professional engagement. Drawing on the proactive motivation model, this study examines whether and how principals' servant leadership exerts cross-level effects on teachers' approach and avoidance job crafting. Data were collected from 1724 teachers nested within 150 kindergartens, and hypotheses were tested using multilevel modeling. The results indicated that principals' servant leadership had significant cross-level effects on teachers' approach and avoidance job crafting, positively predicting approach job crafting and negatively predicting avoidance job crafting. In addition, organizational identification functioned as a cross-level mediator in this relationship, through which servant leadership further enhanced approach job crafting and reduced avoidance job crafting. These findings extend the literature by revealing the motivational pathway linking servant leadership to distinct forms of job crafting and highlight the importance of cultivating a servant leadership climate to foster proactive behaviors among kindergarten teachers.