Journal of Neuroscience Methods最新文献

{"title":"Optimizing stereotaxic injection strategy for AAV-mediated corticospinal tract tracing in mice","authors":"Wei Wang ,&nbsp;Tehan Zhang ,&nbsp;Shaolong Li ,&nbsp;Wenzhao Wang ,&nbsp;Quanhe Jin ,&nbsp;Chi Zhang ,&nbsp;Jie Liu ,&nbsp;Haijian Sun ,&nbsp;Shiqing Feng","doi":"10.1016/j.jneumeth.2026.110679","DOIUrl":"10.1016/j.jneumeth.2026.110679","url":null,"abstract":"<div><h3>Background</h3><div>The corticospinal tract (CST) is a major descending motor pathway essential for voluntary motor control. While adeno-associated virus-mediated anterograde tracing is widely used to label CST projections in mice, what the best stereotaxic injection coordinates and post-injection intervals remain unclear.</div></div><div><h3>New method</h3><div>Here, we systematically evaluated eight cortical injection strategies, differing in anterior-posterior (AP) and medial-lateral (ML) coordinates, the number of injection sites, and post-injection intervals. CST labeling was quantitatively assessed at cervical 2, cervical 5, thoracic 2, thoracic 6, and lumbar 2 spinal levels using transduced axon count (TAC), mean fluorescence intensity (MFI) and transduced area within the dorsal columns, normalized to the C2 segment.</div></div><div><h3>Results</h3><div>Across the tested AP/ML coordinates and single- versus multi-site injections, TAC and MFI were broadly comparable across cervical and thoracic levels, with the exception of reduced L2 labeling in Group IV (AP +0.70 mm). Furthermore, reducing the post-injection interval from four weeks to two weeks did not compromise labeling efficiency.</div></div><div><h3>Comparison with existing methods</h3><div>Conventional CST tracing typically requires multiple injections and ≥ 4-week intervals, increasing complexity and duration. Our optimized single-injection, 2-week protocol achieves comparable labeling fidelity while reducing procedural burden and improving reproducibility.</div></div><div><h3>Conclusion</h3><div>We suggest that effective CST labeling from L2 and rostral segments can be achieved with a single-point injection at AP coordinates between 0.0 and + 0.7 mm (ML fixed at 1.2 mm) or ML coordinates from + 0.7 to + 1.5 mm (AP fixed at 0.0 mm). These results establish a simplified, reproducible strategy for CST tracing.</div></div>","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"428 ","pages":"Article 110679"},"PeriodicalIF":2.3,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145944572","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Rapid invisible frequency tagging (RIFT) with a consumer monitor: A proof-of-concept","authors":"Olaf Dimigen,&nbsp;Ioana Badea,&nbsp;Iarina Simon,&nbsp;Mark M. Span","doi":"10.1016/j.jneumeth.2025.110660","DOIUrl":"10.1016/j.jneumeth.2025.110660","url":null,"abstract":"<div><h3>Background</h3><div>Rapid Invisible Frequency Tagging (RIFT) enables neural frequency tagging at rates above the flicker fusion threshold, eliciting steady-state responses to flicker that is almost imperceptible. While RIFT has proven valuable for studying visuospatial attention, it has so far relied on costly projector systems, typically in combination with magnetoencephalography (MEG). The recent emergence of high-speed organic light-emitting diode (OLED) monitors for consumers suggests that RIFT may also be feasible with much more accessible hardware.</div></div><div><h3>New method</h3><div>Here, we provide a proof-of-concept demonstrating successful RIFT using a consumer-grade 480 Hz OLED monitor in combination with electroencephalography (EEG). We also share practical recommendations for achieving precise stimulus timing at 480 Hz with minimal frame drops.</div></div><div><h3>Results</h3><div>In a central fixation task, participants viewed a tapered disc stimulus flickering either centrally or peripherally. Luminance was modulated sinusoidally at 60 Hz or 64 Hz, frequencies at which the flicker was barely visible. Photodiode recordings confirmed that the monitor delivered accurate frame timing with few dropped frames. Cross-coherence analysis between occipital EEG channels and a photodiode revealed robust, frequency-specific neural tagging responses for central stimuli at both frequencies. In comparison, weaker coherence was observed for 60 Hz peripheral flicker.</div></div><div><h3>Conclusions</h3><div>Our findings demonstrate that RIFT can be reliably implemented using affordable stimulation hardware, a low-density EEG montage, and a minimal processing pipeline. We hope that this lowers barriers to entry, facilitating broader use of RIFT in basic research and in applied settings where cost and portability matter.</div></div>","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"428 ","pages":"Article 110660"},"PeriodicalIF":2.3,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145794079","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Screening channelrhodopsins using robotic intracellular electrophysiology and single cell sequencing","authors":"Samuel Ehrlich ,&nbsp;Alexandra D. VandeLoo ,&nbsp;Mohamed Badawy ,&nbsp;Mercedes M. Gonzalez ,&nbsp;Max Stockslager ,&nbsp;Aimei Yang ,&nbsp;Sapna Sinha ,&nbsp;Shahar Bracha ,&nbsp;Demian Park ,&nbsp;Benjamin Magondu ,&nbsp;Bo Yang ,&nbsp;Edward S. Boyden ,&nbsp;Craig R. Forest","doi":"10.1016/j.jneumeth.2025.110663","DOIUrl":"10.1016/j.jneumeth.2025.110663","url":null,"abstract":"<div><h3>Background:</h3><div>Our ability to engineer opsins is limited by an incomplete understanding of how sequence variations influence function. The vastness of opsin sequence space makes systematic exploration difficult.</div></div><div><h3>New method:</h3><div>In recognition of the need for datasets linking opsin genetic sequence to function, we pursued a novel method for screening channelrhodopsins to obtain these datasets. In this method, we integrate advances in robotic intracellular electrophysiology (<u>P</u>atch) to measure optogenetic properties (<u>E</u>xcite), harvest individual cells of interest (<u>P</u>ick) and subsequently sequence them (<u>S</u>equence), thus tying sequence to function.</div></div><div><h3>Results:</h3><div>We used this method to sequence more than 50 cells with associated functional characterization. We further demonstrate the utility of this method with experiments on heterogeneous populations of known opsins and single point mutations of a known opsin. Of these point mutations, we found C160W ablates ChrimsonR’s response to light.</div></div><div><h3>Conclusion and comparison to existing methods:</h3><div>Compared to traditional manual patch clamp screening, which is labor-intensive and low-throughput, this approach enables more efficient, standardized, and scalable characterization of large opsin libraries. This method can enable opsin engineering with large datasets to increase our understanding of opsin sequence–function relationships.</div></div>","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"428 ","pages":"Article 110663"},"PeriodicalIF":2.3,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145800610","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Feature library for behavioural characterization of early and late seizures in an experimental model of post-traumatic epilepsy","authors":"Pedro Andrade,&nbsp;Asla Pitkänen","doi":"10.1016/j.jneumeth.2025.110671","DOIUrl":"10.1016/j.jneumeth.2025.110671","url":null,"abstract":"<div><h3>Background</h3><div>To generate a behavioral feature library for a more granular description of seizure semiology in rats with traumatic brain injury (TBI). To compare the new approach to the Racine score.</div></div><div><h3>New method</h3><div>A library of 59 seizure-related behavioral features was generated by annotating 329 seizures in 31 rats with TBI, which were monitored using high-resolution video-electroencephalogram. Of the 329 seizures, 149 were early, 85 post-electrode implantation (6th post-injury month), and 95 late seizures (7th post-injury month). Of the 59 behavioral features, 3 were pre-ictal, 43 ictal, and 13 post-ictal. Of the 43 ictal features, 7 related to consciousness, 5 to mouth and whiskers, 2 to eyes, 7 to head, 2 to ears, 6 to paws, 12 to body and tail, 2 to autonomic function, and 1 to wet-dog shakes.</div></div><div><h3>Results</h3><div>Early, post-implantation, and late seizures showed different behavioral phenotypes (p &lt; 0.001). The number of behavioral features in post-electrode implantation and late seizures was greater than that in early seizures (p &lt; 0.05). Behavioral features did not reliably differentiate transitions from pre-ictal to ictal or from ictal to post-ictal phases.</div></div><div><h3>Comparison with existing methods</h3><div>Ninety-one percent of early, 45 % of post-electrode implantation, and 18 % of late seizures with a Racine score of 0 showed up to 6–7 ictal-related behaviors.</div></div><div><h3>Conclusions</h3><div>The Proposed feature list can be applied for the harmonization of data analysis and reporting, and training of video-based seizure detection algorithms to speed up non-invasive, affordable epilepsy diagnosis and assessment of treatment effects in TBI models.</div></div>","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"428 ","pages":"Article 110671"},"PeriodicalIF":2.3,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145906204","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A causal attention network with time frequency channel feature fusion for epileptic seizure prediction","authors":"Yimin Qu ,&nbsp;Songhui Rao ,&nbsp;Ting Li ,&nbsp;Ying Li ,&nbsp;Yan Niu ,&nbsp;Ruiyun Chang ,&nbsp;Xianchuan Chen ,&nbsp;Bin Wang","doi":"10.1016/j.jneumeth.2025.110665","DOIUrl":"10.1016/j.jneumeth.2025.110665","url":null,"abstract":"<div><h3>Background:</h3><div>Epilepsy poses ongoing physical and mental threats and causes substantial economic burdens. Better seizure forecasting enables faster medical responses, improving patients’ quality of life and lowering healthcare costs. Research mainly focuses on early forecasting within a short preictal window, often too brief for effective drug administration. A major challenge is that a longer preictal phase may resemble the interictal state, making differentiation difficult.</div></div><div><h3>New methods:</h3><div>We propose a causal attention network (CANet) with a longer interictal and preictal of 1 h and 2 h respectively as the research object. In the feature extraction, a dilated causal convolution network is employed to extract local features. Causal attention is innovatively incorporated into epilepsy prediction to capture global correlation features. The complementary integration of these two methods enhances feature extraction and enables a more precise distinction between interictal and preictal periods. A double-layer dynamic window algorithm is developed for seizure prediction.</div></div><div><h3>Results:</h3><div>We evaluate the performance on Freiburg and CHB-MIT datasets. On the Freiburg dataset, the sensitivity(Sen) of the 1/2-hour preictal intervals was 100.00%/96.67%, with a false alarm rate per hour (FAR) of 0.0077/h/0.0472/h, and the average prediction time (APT) was 97.59 min. On the CHB-MIT dataset, we achieved Sen of 97.06%/92.31%, FAR of 0.0251/h/0.0666/h, and APT of 94.85 min, under the same conditions.</div></div><div><h3>Comparison with existing methods and conclusion:</h3><div>Our approach outperforms most of the previous methods, and the intracranial EEG (Freiburg) can more effectively distinguish interictal and preictal periods than scalp EEG (CHB-MIT).</div></div>","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"428 ","pages":"Article 110665"},"PeriodicalIF":2.3,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145928232","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Long term cultures of Xenopus and Drosophila neurons and glial cells","authors":"Noémie Frère,&nbsp;Lydia Chelbi,&nbsp;Louise Mathé,&nbsp;Catherine Lubetzki,&nbsp;Bruno Stankoff,&nbsp;Bernard Zalc","doi":"10.1016/j.jneumeth.2025.110649","DOIUrl":"10.1016/j.jneumeth.2025.110649","url":null,"abstract":"<div><h3>Background</h3><div><em>In vitro</em> models are crucial for exploring cellular and molecular mechanisms underlying central nervous system (CNS) development and function. While long-term neuron–glia cultures have been well established in rodents, fewer approaches exist for amphibians or invertebrates.</div></div><div><h3>New method</h3><div>Here, we present standardized protocols for the dissociation and culture of CNS cells from <em>Xenopus laevis</em> tadpoles (stages 40–46) and <em>Drosophila melanogaster</em> L3 larvae. We developed species-specific media enabling the survival and differentiation of mixed neuronal and glial populations for at least 14 days.</div></div><div><h3>Results</h3><div>Immunofluorescence analyses revealed progressive neuronal network formation and glial maturation, including oligodendrocyte process extension and early signs of myelination in Xenopus, and stable glial subtypes interacting with neurons in Drosophila.</div></div><div><h3>Comparison with existing methods</h3><div>Despite an initial cell loss during the first week, cultures stabilized thereafter, maintaining representative populations of neurons, astrocytes, oligodendrocytes, and microglia in Xenopus, and neurons with ensheathing, wrapping and cortical glia in Drosophila. Cell populations at D14 are similar to those found in more commonly used cultures.</div></div><div><h3>Conclusion</h3><div>This work provides robust platforms for investigating neuron–glia interactions in both vertebrate and invertebrate models.</div></div>","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"427 ","pages":"Article 110649"},"PeriodicalIF":2.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145687497","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Stereotactic intracranial implantation of patient-derived glioblastoma cells in rats: A xenograft modeling approach","authors":"Faezeh Yaghoubi ,&nbsp;Safieh Ebrahimi ,&nbsp;Ali Gorji ,&nbsp;Maryam Khaleghi Ghadiri","doi":"10.1016/j.jneumeth.2025.110658","DOIUrl":"10.1016/j.jneumeth.2025.110658","url":null,"abstract":"<div><h3>Background</h3><div>Glioblastoma (GBM) is an aggressive primary brain cancer with a poor prognosis. Preclinical animal models are essential for studying GBM pathophysiology and therapy; however, existing models often fail to fully capture the tumor's heterogeneity and the partially immunodeficient microenvironment critical to its progression.</div></div><div><h3>New method</h3><div>We developed a novel GBM model using the stereotactic implantation of primary patient-derived GBM cells from various donors into the brains of rats subjected to transient, partial immune suppression.</div></div><div><h3>Results</h3><div>This model combines the biological heterogeneity of patient-derived cells with the anatomical advantages of a larger rodent brain, improving spatial tumor localization. The use of a mixed primary cell population better recapitulates human GBM heterogeneity. Furthermore, the model's partially preserved immune environment allows for the investigation of tumor-immune interactions.</div></div><div><h3>Comparison with existing methods</h3><div>Unlike fully immunodeficient models, our approach maintains a more physiologically relevant, partially intact immune system. Compared to murine models, the rat brain offers superior spatial resolution for tumor analysis and intervention.</div></div><div><h3>Conclusions</h3><div>This method provides a reliable and translational platform that enhances the fidelity of preclinical GBM research. It offers an improved tool for drug evaluation and the development of personalized therapeutic strategies by more accurately mimicking the complex and heterogeneous nature of human GBM.</div></div>","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"427 ","pages":"Article 110658"},"PeriodicalIF":2.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145756843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"REST is a superior EEG re-referencing method: A commentary on the rCAR by Kyle Q. Lepagea et al.","authors":"Yutong Yao ,&nbsp;Xinru Lan ,&nbsp;Cheng Luo ,&nbsp;Li Dong","doi":"10.1016/j.jneumeth.2025.110659","DOIUrl":"10.1016/j.jneumeth.2025.110659","url":null,"abstract":"","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"427 ","pages":"Article 110659"},"PeriodicalIF":2.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145762939","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Propagation mapping using iterative independent component analysis for seizure onset zone localization in temporal lobe epilepsy","authors":"Bingyang Cai ,&nbsp;Shize Jiang ,&nbsp;Jiwei Li ,&nbsp;Zhengwei Zhong ,&nbsp;Haiqing Zhang ,&nbsp;Zengji Zhang ,&nbsp;Xiaolai Ye ,&nbsp;Weiqi Bao ,&nbsp;Jie Hu ,&nbsp;Liang Chen ,&nbsp;Xiaoying Liu ,&nbsp;Jie Luo","doi":"10.1016/j.jneumeth.2025.110648","DOIUrl":"10.1016/j.jneumeth.2025.110648","url":null,"abstract":"<div><h3>Background</h3><div>Epilepsy affects approximately 70 million people worldwide, with a third of them being drug-resistant and requiring surgical intervention. Accurate localization of the seizure onset zone (SOZ) is crucial for effective surgery but remains challenging.</div></div><div><h3>New method</h3><div>We proposed a method using iterative independent component analysis (ICA) to map seizure propagation of drug-resistant temporal lobe epilepsy (TLE). For each assumed seizure origin, ICA was applied to the remaining contacts to identify propagation components, with the highest correlating component being the propagating signal. Iterative removal of nearby contacts revealed spatial propagation profiles. Machine learning models were applied to the propagation profiles to distinguish the SOZ.</div></div><div><h3>Results</h3><div>Seizure propagation features differed significantly between SOZ and non-SOZ contacts in seizure free patients (both local cohort N = 21, and independent dataset N = 13), but not in non-seizure free patients (N = 11). Propagation-based classifiers achieved robust performance (AUC = 0.85), outperforming iEEG source imaging (AUC = 0.73). In mesial TLE, propagation maintained high accuracy (AUC = 0.84) while iEEG source imaging dropped markedly (AUC = 0.64).</div></div><div><h3>Comparison with existing methods</h3><div>Traditional methods for SOZ localization, such as visual inspection of SEEG and source imaging techniques, rely heavily on expert interpretation. iEEG source imaging assumes linear forward models and can be susceptible to inaccuracies due to electrode placement and noise. In contrast, our proposed iterative ICA approach is purely data-driven and adaptively identifies the dominant propagation pathways across individual seizures.</div></div><div><h3>Conclusion</h3><div>This work introduces a data-driven strategy to characterize seizure propagation, potentially improving SOZ localization with deep brain origins.</div></div>","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"427 ","pages":"Article 110648"},"PeriodicalIF":2.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145658764","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A multi-scale deep CNN based on attention mechanism for EEG emotion recognition","authors":"Yi Zhou,&nbsp;Ruiwen Jiang,&nbsp;Jingxiang Zhang","doi":"10.1016/j.jneumeth.2025.110662","DOIUrl":"10.1016/j.jneumeth.2025.110662","url":null,"abstract":"<div><h3>Background</h3><div>Recognizing emotion is a crucial project within the domain of brain-computer interface technology. Recently, researchers have found that deep learning have been proven to be superior to machine learning, but how to obtain more discriminative features still faces great challenges.</div></div><div><h3>New method</h3><div>We propose a multi-scale convolutional neural network (MSCNN) based on channel attention and spatial attention (CSA-MSCNN) for EEG emotion recognition. The channel attention enhances the feature extraction ability of critical channels by generating channel weights, while suppressing noise or interference from redundant channels. The spatial attention helps the model to more precisely locate key areas related to emotion by generating a spatial weight matrix. To extract more comprehensive features, CSA-MSCNN uses MSCNN for feature extraction, with smaller convolutional kernels capturing the local details of the signals, and larger convolutional kernels with a broader receptive field to obtain deeper signal information.</div></div><div><h3>Results</h3><div>CSA-MSCNN achieves average accuracies of 95.75 % and 95.39 % for three-class classification of valence and arousal on DEAP, respectively, while achieving an average three-class classification accuracy of 90.48 % on SEED.</div></div><div><h3>Comparison with existing methods</h3><div>The classification accuracy of CSA-MSCNN is not only significantly better than traditional machine learning models, but also shows strong competitiveness compared with mainstream deep learning models such as graph convolutional neural network (GCNN).</div></div><div><h3>Conclusions</h3><div>CSA-MSCNN addresses the issues of multiple EEG signal channels and complex regional information.</div></div>","PeriodicalId":16415,"journal":{"name":"Journal of Neuroscience Methods","volume":"427 ","pages":"Article 110662"},"PeriodicalIF":2.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145794043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}