发表机构
Southeast University（东南大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文以DGCNN在SEED、SEED-IV上的实验为例，分离EEG情绪识别的评估要素，发现跨被试泛化差距非简单欠拟合导致，需区分不同评估结果的意义。
AI 中文摘要
脑电图（EEG）情绪识别的报告准确率取决于完整的评估流程，而非仅分类器本身。本文将目标指标、开发流程和报告规则分离，采用存档的动态图卷积神经网络（DGCNN）路径在SEED和SEED-IV数据集上开展说明性案例研究。在协议匹配的被试依赖检验中，SEED结果与公开参考值的差值在1.47个百分点以内；SEED-IV的3.40点差值仍未解决。在30条匹配的SEED被试-会话轨迹中，基于重复测试集评估的检查点选择将平均窗口准确率从第80轮的0.7855提升至0.8892。在五折被试不重叠评估下，验证集选择的检查点在SEED上达到训练参与者试次准确率0.9990，在SEED-IV上为0.9920；完全保留参与者的准确率在SEED上为0.5348（95%条件被试水平偏差校正加速[BCa]区间[0.4667, 0.5985]），SEED-IV估计值为0.3954（[0.3343, 0.4648]），因协议匹配兼容性检验未解决，仅作为次要敏感性证据报告。观察到的训练到保留被试的差距与简单优化欠拟合不一致，但无法将被试身份与实现、预处理、表示或分布因素分离。支持性分析进一步显示，参与者排名取决于表示和时间尺度，而开发选择的尾部风险集成在单独的最终评估中未建立正增益。因此，应将被试依赖、被试不重叠和跨会话结果作为不同问题的答案报告。
英文摘要
Checkpoint selection can improve an electroencephalography (EEG) emotion-recognition score without improving performance on other trials. We compared selection and scoring on disjoint trial pools along fixed training trajectories. A same-session SEED study comprised 300 trajectories from 15 participants, two models and five trial-role rotations. Another 276 trajectories extended the comparison to separate training, validation and target sessions in SEED, SEED-IV and SEED-V, with 15, 15 and 16 participants, respectively. Increasing the candidate set from five to 80 checkpoints raised the same-session dynamical graph convolutional neural network (DGCNN) selection-pool score by 6.24 percentage points, while the other-pool change was -1.24 (descriptive 95% participant-bootstrap interval -2.67 to 0.26). Under the cross-session design, DGCNN retained gains of 4.47 [1.79, 7.25], 4.27 [2.00, 6.53] and 2.75 [0.79, 4.78] points across the three datasets. Multilayer perceptron results varied across datasets. Common-time-range analysis and trial-level scoring supported the respective patterns. Both target pools supplied labels for symmetric checkpoint selection; these results concern retention across trials, not evaluation without target labels. Complete policy curves and executable reconstruction support reporting selected-score improvements alongside performance on other trials, rather than assuming a universal penalty for checkpoint search.
Comments15 pages, 3 figures, 3 tables; 27-page supplementary information and reproducibility materials in ancillary files. Substantially revised title and analysis, added SEED-IV and SEED-V experiments, and updated author list. Submitted to Neuroinformatics
