【摘要】 目前主流的精神分裂症识别方法普遍依赖手动特征提取与静态脑网络构建，难以充分挖掘出EEG脑电信号下的多尺度动态特征与潜在脑区连接关系，且多未在特征提取时做到时空间特征分离，导致模型耦合度过高.为此，本文提出一种基于多尺度门控卷积图卷积网络的分类模型，在任务态精分数据集中进行实验.本模型在Theta频段与相位锁定值连接指标条件下平均准确率为95.51%，其余多项分类指标均优于现有方法，证实了该结构在性能上的优越性，为精神分裂症的客观辅助诊断提供了新思路.更多还原

【Abstract】 Current mainstream methods for schizophrenia recognition largely rely on manual feature extraction and static brain network construction,which makes it difficult to fully capture the multi-scale dynamic characteristics of EEG signals and the potential inter-regional brain connectivity.Moreover,most existing approaches fail to achieve spatiotemporal separation during feature extraction,leading to excessive coupling in the model.To address these limitations,this paper proposes a classification model based on a multi-scale gated convolutional graph convolutional networkand conducts experiments on task-related schizophrenia datasets. Under the Theta frequency band and the phaselocking valueconnectivity measure,the proposed model achieves an average accuracy of 95.51%,and outperforms existing methods across multiple evaluation metrics.These results demonstrate the superiority of the proposed structure and provide a novel perspective for objective auxiliary diagnosis of schizophrenia.更多还原