International journal of neural systems最新文献

{"title":"Robust and Efficient Neural Architecture Design via Implicit Euler Explicitization.","authors":"Yue Huang, Jianwei Zhang, Lei Zhang, Yan Wang, Jie Wu, Zhanji Shi, Rui Gao, Yilin Guo, Chang Liu, Chaobang Gao","doi":"10.1142/S0129065727500146","DOIUrl":"10.1142/S0129065727500146","url":null,"abstract":"<p><p>Modeling deep neural networks as continuous-time dynamical systems has emerged as a powerful approach for analyzing their evolution and robustness. By leveraging the superior numerical stability of implicit methods, several studies have proposed the implicit Euler-based architecture, which significantly enhances both robustness and generalization under adversarial attack. However, existing methods require solving implicit equations iteratively, which substantially increases the computational cost. To overcome these limitations, this paper proposes an explicit reformulation of the implicit Euler method, termed <b>I</b>mplicit-to-<b>EX</b>plicit ResNet (IEX-ResNet). IEX-ResNet preserves the inherent numerical stability of implicit methods while eliminating the need for iterative solvers, thereby enabling the efficient training of robust neural models. Theoretical analysis demonstrates that IEX-ResNet, under local linearization and spectral radius constraint, achieves Lipschitz-bounded perturbation propagation and maintains a stable approximation of the underlying implicit dynamics. Experimental results show that, compared to conventional implicit Euler architectures, IEX-ResNet achieves improved robustness across diverse challenging settings and maintains competitive robustness while reducing training time by nearly 50% under standard evaluation. Moreover, IEX-ResNet provides practical benefits in a real-world document image restoration task. These results suggest that implicit Euler explicitization is a practical route toward efficient robust neural architecture design.</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":" ","pages":"2750014"},"PeriodicalIF":6.4,"publicationDate":"2027-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148803605","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Compensation-Balanced Recurrent Neural Architecture Based on Nonlinear Spiking Neural Membrane Systems.","authors":"Jun Fu, Hong Peng, Bing Li, Ziyin Zhou","doi":"10.1142/S0129065727500213","DOIUrl":"10.1142/S0129065727500213","url":null,"abstract":"<p><p>Nonlinear spiking neural P (NSNP) systems offer a biologically inspired framework for modeling nonlinear temporal dynamics via spike consumption and generation. Existing NSNP-based recurrent architectures, such as the long short-term memory model inspired from spiking neural P systems (LSTM-SNP), primarily rely on implicit gating mechanisms and lack explicit state correction during recurrent propagation, which can lead to unstable state evolution and excessive information decay in long-term temporal modeling. To address this limitation, a compensation-balanced LSTM-SNP (CBLSTM-SNP) architecture is proposed. The model introduces a compensation-driven regulation mechanism, comprising a regulation branch and an adaptive compensation branch, to dynamically adjust hidden state evolution. From a nonlinear dynamical systems perspective, CBLSTM-SNP establishes a balanced recurrent state transition integrating memory preservation, nonlinear spike processing, and adaptive state restoration. Theoretical analysis confirms the boundedness and Lipschitz continuity of the proposed recurrent dynamics. Empirical evaluations on five benchmark time series datasets demonstrate that CBLSTM-SNP produces smoother hidden state trajectories and achieves comparable forecasting performance compared with several representative recurrent and hybrid models.</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":" ","pages":"2750021"},"PeriodicalIF":6.4,"publicationDate":"2027-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148868510","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multidimensional Microstate Alterations and Interpretable Classification of Electrical Status Epilepticus During Sleep Using Multi-Channel EEG.","authors":"Shishi Tang, Zhipeng He, Yicong Liu, Xinxin Peng, Yuxuan Li, Yuting An, Jianping Man, Gan Sen, Guanzhong Ni, Ziyi Chen, Yi Zhou","doi":"10.1142/S0129065727500122","DOIUrl":"10.1142/S0129065727500122","url":null,"abstract":"<p><p>Electrical status epilepticus in sleep (ESES) is an epilepsy syndrome requiring early diagnosis. Its diagnostic subjectivity and therapeutic inadequacy necessitate the urgent discovery of objective markers. This study developed a microstate-centered multidimensional EEG framework to distinguish epilepsy patients with ESES from those without ESES and to identify interpretable neurophysiological alterations. Broadband spatiotemporal parameters, frequency-specific microstate features, nonlinear complexity, symbolic-sequence dynamics, Spike-Wave Index (SWI) correlations, and exploratory classification performance were evaluated. ESES-related alterations were concentrated in microstate B and C, involving prolonged and more variable temporal persistence, altered occurrence frequency, increased global field strength, frequency-sensitive duration changes, and reduced frequency chaos game representation (FCGR) entropy. Markov order features were positively associated with SWI-related electrophysiological burden, suggesting potential severity-related sequence dynamics. A five-feature L2-regularized logistic regression model achieved promising exploratory discrimination, with [Formula: see text] and [Formula: see text]. This study indicated that ESES is associated with more constrained and less flexible microstate dynamics, while a compact, interpretable microstate-derived model achieved promising exploratory discrimination. These findings may provide novel analytical perspectives for deciphering ESES pathophysiology and identifying candidate neurophysiological markers for future assessment.</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":" ","pages":"2750012"},"PeriodicalIF":6.4,"publicationDate":"2027-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148803633","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Semantic-Aware Semi-Supervised Systematic Segmentation for Brain Magnetic Resonance Image.","authors":"Hussain Ahmad Madni, Silvia Zottin, Axel De Nardin, Rao Muhammad Umer, Gian Luca Foresti","doi":"10.1142/S0129065727500183","DOIUrl":"10.1142/S0129065727500183","url":null,"abstract":"<p><p>Accurate segmentation of brain MRI images is essential for early diagnosis and prevention of brain tumors and strokes. However, it is challenging due to data scarcity, domain shift, complexity, and blurred edges. Recent methods use mean-teacher model learning for the semantic details of an individual image based on the efficiency of teacher architecture. However, to enhance the segmentation in brain MRI, the paper introduces SemS4, a semi-supervised method on the basis of cross-image architecture. SemS4 uses a single architecture integrating an Edge Prototype Attention (EPA) and a foreground prototype attention (FPA). More specifically, EPA uses edge prototype integrated with adaptive edge container to improve the edge features for which global edge features are stored to produce edge prototype. Moreover, FPA stabilizes the segmentation by transferring complementary foreground information. Furthermore, the proposed method uses Pixel Affinity Loss (PAL) for missing contextual correlation during supervision to enhance segmentation performance on edges. The performance of the proposed approach in experimental results for two brain MRI benchmarks, LGG and BRISC, validates the superiority over existing methods under different partitioning settings.</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":" ","pages":"2750018"},"PeriodicalIF":6.4,"publicationDate":"2027-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148868508","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Neuroimaging Biomarkers for Brain Disorder Diagnosis Using Advanced Learning Techniques.","authors":"Iulia-Andreea Ion, Camelia Chira","doi":"10.1142/S0129065727500158","DOIUrl":"10.1142/S0129065727500158","url":null,"abstract":"<p><p>Mental illness is becoming increasingly common in everyday life, affecting a growing number of people. Functional Magnetic Resonance Imaging (fMRI) is an effective technique for detecting mental disorders. Resting-state fMRI (rs-fMRI) analyses spontaneous low-frequency oscillations in the Blood Oxygen Level Dependent signal to examine the functional architecture of the brain. fMRI and rs-fMRI data processing contribute to mapping neural activity in specific brain regions, enabling accurate localization of cognitive functions and improving our understanding of Functional Connectivity and interactions among brain regions. This paper focuses on the use of Machine Learning and Deep Learning techniques for fMRI analysis to identify brain features that may be associated with Bipolar Disorder (BD) and Alzheimer's Disease (AD). The first step in the proposed approach is the preprocessing of fMRI and rs-fMRI data using two distinct methods: the first combines Statistical Parametric Mapping and the CONN toolbox for improved results, and the second focuses on the MELODIC software. The proposed approach begins with fMRI and rs-fMRI preprocessing using two pipelines: one integrating Statistical Parametric Mapping (SPM) with the CONN toolbox and another based on the MELODIC framework. The aim is to evaluate how preprocessing choices affect functional connectivity representations and classification performance when combined with classical Machine Learning (ML) methods, namely Support Vector Machines and Random Forests, and the Deep Learning (DL) model AlexNet. Given the altered brain connectivity patterns observed in Bipolar Disorder and Alzheimer's Disease, these methods are employed to assess preprocessing-learning interactions across both disorders. Experimental results emphasize that Connbox-SPM preprocessing combined with AlexNet yields the best results for both mental disorders, with Bipolar Disorder reporting 78.2% accuracy and AD achieving 82.4% accuracy.</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":" ","pages":"2750015"},"PeriodicalIF":6.4,"publicationDate":"2027-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149043068","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep Model Families for EEG-Based Multi-Class Dementia Classification.","authors":"Soner Kotan, Aydin Akan","doi":"10.1142/S0129065727500055","DOIUrl":"10.1142/S0129065727500055","url":null,"abstract":"<p><p>Deep learning (DL) has shown considerable promise for EEG-based dementia assessment; however, rigorous cross-family comparisons under leakage-free and clinically meaningful evaluation protocols remain limited. To address this gap, we benchmarked twelve DL architectures from four major families - recurrent neural networks (RNNs), temporal convolutional networks (TCNs), Transformers, and self-supervised learning (SSL) - for the multi-class differentiation of Alzheimer's disease (AD), frontotemporal dementia (FTD), and healthy controls (HC). The analysis was conducted on a clinically acquired EEG dataset comprising 36 AD patients, 23 FTD patients, and 29[Formula: see text]HC participants. To prevent subject-level data leakage, we employed strict subject-independent five-fold cross-validation, and to improve clinical relevance, final predictions were aggregated at the subject level using majority voting across EEG segments. Model performance was evaluated using classification metrics, learning stability, statistical ranking, computational efficiency, and interpretability analysis. Among all benchmarked architectures, the self-supervised SimCLR-Time framework achieved the strongest subject-level performance, reaching a mean accuracy of [Formula: see text] and a weighted <i>F</i>1-score of [Formula: see text]. The time-series transformer (TST) was the best-performing fully supervised model, with an accuracy of [Formula: see text], followed closely by the hybrid LSTM-FCN model at [Formula: see text]. In contrast, the EEG-Seq2Seq model showed the weakest performance ([Formula: see text]). Computational profiling further showed that TCN-Attention offered a favorable trade-off between accuracy and edge-oriented efficiency, while SimCLR-Time combined strong diagnostic performance with the fastest inference latency. In addition, Integrated Gradients-based saliency analysis suggested that the top-performing model relied on structured spatiotemporal EEG patterns rather than arbitrary temporal artifacts. Overall, these findings provide a controlled within-dataset benchmark for multi-class dementia classification from raw EEG and suggest that self-supervised and attention-based architectures are promising directions for clinically oriented EEG decision-support systems. However, external validation on larger and more heterogeneous cohorts remains necessary before broader generalization can be claimed.</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":" ","pages":"2750005"},"PeriodicalIF":6.4,"publicationDate":"2027-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148457929","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Large Pre-trained EEG Model for Electrical Status Epilepticus during Sleep Detection.","authors":"Yuxuan Li, Zhipeng He, Shishi Tang, Rui Yang, Jianping Man, Yi Zhou, Jiahao Zhu, Xinxin Peng, Yicong Liu, Ziyi Chen","doi":"10.1142/S0129065727500109","DOIUrl":"10.1142/S0129065727500109","url":null,"abstract":"<p><p>Electrical status epilepticus during sleep (ESES) is a severe but often underdiagnosed epileptic encephalopathy that can significantly impair cognitive development in children. Current diagnostic practices rely heavily on expert visual inspection of spike-and-slow-wave complexes in electroencephalogram (EEG) recordings. This process is time-consuming, labor-intensive, and subjective, often resulting in poor inter-rater agreement. To address these limitations, this paper introduces a novel ESES automatic detection framework based on a large pre-trained EEG model. The framework employs a pre-trained NeuroEncoder to extract temporal and frequency features from spike-and-slow-wave complexes, and uniquely incorporates a temporal and spatial embedding module to enhance the model's ability to recognize spatiotemporal patterns. By leveraging the multi-head attention mechanism of stacked Transformers, the framework further captures complex relationships and deep features. A symmetric label smoothing cross-entropy is introduced, which applies both forward and backward constraints to reduce overfitting to noisy labels and improve robustness to label noise and individual variability. The proposed framework can automatically quantify spike-and-slow-wave complexes and calculate the spike-wave index (SWI) for ESES diagnosis. Evaluation was conducted on EEG recordings from 35 patients using a leave-one-subject-out cross-subject strategy. The results demonstrate excellent performance, with an accuracy of 91.91%, sensitivity of 91.52%, and specificity of 92.37%. The high agreement between the model and expert annotation is confirmed by a Kappa coefficient of 0.8368 and Gwet's AC1 of 0.8394. The proposed method demonstrates superior generalization across subjects and outperforms existing approaches. The proposed framework provides an efficient and reliable tool for automatic ESES detection, supporting early intervention and improved clinical management for patients with ESES.</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":" ","pages":"2750010"},"PeriodicalIF":6.4,"publicationDate":"2027-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148762154","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhancing Epileptic Seizure Identification by Exploring Swin Transformer Integration within Conformer Architecture.","authors":"Rufeng Wen, Jian Lian","doi":"10.1142/S0129065727500079","DOIUrl":"10.1142/S0129065727500079","url":null,"abstract":"<p><p>This study proposes a novel enhancement to the Conformer architecture for epileptic seizure identification by replacing the standard Vision Transformer (ViT) with the Swin Transformer. The proposed Swin-Conformer model leverages the hierarchical patch merging and shifted-window self-attention mechanisms of the Swin Transformer to better capture both local electrophysiological patterns and long-range temporal dependencies in electroencephalogram (EEG) signals. To handle the significant class imbalance inherent in epileptic EEG data, a weighted focal loss function is employed during training. The model is evaluated on the CHB-MIT dataset through stratified 10-fold cross-validation and validated on the independent Bonn EEG dataset. Experimental results demonstrate that the proposed model achieves accuracy of 99.24%, specificity of 99.55%, and sensitivity of 98.47% on the segment-based evaluation, as well as an event-based sensitivity of 99.50%, outperforming the original Conformer and state-of-the-art baseline methods. Ablation studies confirm that the performance gains originate from the Swin Transformer's hierarchical multi-scale feature representation and efficient local-global attention mechanism. Statistical analysis confirms that the improvements are statistically significant ([Formula: see text]).</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":" ","pages":"2750007"},"PeriodicalIF":6.4,"publicationDate":"2027-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148674882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Dual Soft Attention Network for Automatic Sleep Staging via Single-Lead Electrocardiogram Signals.","authors":"Xuanhao Qi, Qinying Li, Hongze Liu, Yinlin Cheng, Bingo Wing-Kuen Ling, Zhensheng Hu","doi":"10.1142/S0129065727500110","DOIUrl":"10.1142/S0129065727500110","url":null,"abstract":"<p><p>Polysomnography (PSG) is the current gold standard for sleep monitoring; however, it is complex, burdensome, and costly. In contrast, commercial health monitoring devices, such as the Apple Watch, offer sleep scoring at a low cost over a few days. This paper proposes a dual-attention mechanism-based sleep staging method using single-lead electrocardiogram (ECG) signals, named SleepECGNet. The method extracts multi-scale features by employing convolutional kernels of different sizes, thereby capturing subtle details in the ECG signals. The model then integrates an SE module and a Luong attention-based encoder-decoder architecture to filter redundant information and dynamically focus on different sleep stages. Experiments were conducted using the HMC, MGH, and SHHS1 datasets for sleep staging tasks and the Apnea-ECG corpus was employed to execute transfer learning strategies directed at sleep apnea recognition. Comprehensive ablation experiments were conducted to verify the effectiveness of each component. Moreover, an analysis of sequence duration demonstrated that expanding the input timeframe directly improves classification outcomes.</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":" ","pages":"2750011"},"PeriodicalIF":6.4,"publicationDate":"2027-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148914193","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Functional Connectivity Among Prefrontal, Hippocampal, and Temporal Systems Supporting Conceptual Change in Science Learning.","authors":"Hsiao-Ching She, Meng-Jun Chen, Hsin-Ju Tsai, Jeng-Ren Duann, Li-Yu Huang, Wen-Chi Chou, Ching-Ying Hsueh","doi":"10.1142/S0129065727500134","DOIUrl":"https://doi.org/10.1142/S0129065727500134","url":null,"abstract":"<p><p>Scientific conceptual change poses a cognitive challenge in science learning, requiring cognitive conflict resolution and restructuring of entrenched semantic knowledge. Although neuroscience implicates coordinated prefrontal cognitive control and hippocampal-temporal systems in conflict resolution and semantic memory updating, direct neural evidence of their interaction during conceptual change remains limited. Using functional MRI and Granger Causality (GC) analyses, we examined neural mechanisms distinguishing successful from unsuccessful conceptual change across task complexities in 48 undergraduate and graduate STEM majors who completed 39 conceptual change tasks (19 weak restructuring task, 20 radical restructuring task). Learners exhibited greater conceptual change success during weak than radical restructuring, with enhanced BOLD activation in the superior, middle, and inferior frontal gyri (SFG, MFG, IFG), Hippocampus (HIP), and Superior Temporal Gyrus (STG) during successful versus unsuccessful conceptual change. GC analysis identified a minimal network (SFG, MFG→HIP, STG→MTG, MTG→HIP) common to both conditions, whereas successful conceptual change engaged a denser prefrontal-hippocampal-temporal network, marked by additional unidirectional (MFG→SFG, MFG→IFG, IFG→HIP) and bidirectional (MFG↔STG, STG↔HIP, MTG↔HIP) connectivity. Unsuccessful conceptual change was characterized by shallow, fragmented connectivity and limited cross-regional integration. Collectively, these findings suggest that effective conceptual change arises from coordinated neural networks integrating executive control, semantic processing, and memory updating.</p>","PeriodicalId":94052,"journal":{"name":"International journal of neural systems","volume":"37 2","pages":"2750013"},"PeriodicalIF":6.4,"publicationDate":"2027-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149187017","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}