发表机构
Inclusive Brains; Wavestone; Human Technology Foundation; Mohamed bin Zayed University of Artificial Intelligence; Biotech Dental Group; IBM Technology; IBM France Lab（Inclusive Brains; 威斯顿; 人类技术基金会; 穆罕默德·本·扎耶德人工智能大学; 生物科技牙科集团; IBM技术公司; IBM法国实验室）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究脑电图运动想象解码中个体差异问题，通过三个公共数据集进行大规模基准测试，比较多种方法，发现受试者水平存在显著异质性，利用基准构建紧凑流程组合，Top-K均值给出最佳权衡，可利用异质性实现个性化。
AI 中文摘要
强大的个体间变异性限制了稳健的脑电图运动想象解码，难以识别适用于所有用户的流程。我们在三个公共数据集（Cho2017的52名受试者、PhysionetMI的109名受试者和Zhou2016的4名受试者）上进行了大规模、标准化的会话内解码流程基准测试。使用通用的MOABB左-右想象设置、两个频段（8-15Hz和8-30Hz）以及特征提取、预处理和分类步骤的广泛组合，分析了216,714个原始评估行，结构化聚合后分别产生44,928、109,000和4,192个受试者水平的观测值。协方差切空间投影（cov-tgsp）和共同空间模式（CSP）始终定义了最强的方法家族，但其相对排序取决于数据集。在Cho2017上，最佳家族水平平均准确率来自8-30Hz的cov-tgsp（0.712±0.140），而Zhou2016则青睐CSP（8-15Hz时为0.832±0.121）。这些总体排名掩盖了受试者水平的显著异质性。然后，我们将基准用作构建大小为K的紧凑流程组合的实证性能景观，比较了几种构建程序，包括基于排名的Top-K均值启发式方法和基于搜索的策略，结果大致一致，Top-K均值给出了最佳权衡。因此，这种景观取决于受试者，并且可以通过紧凑组合利用这种异质性，使个性化更可行。
英文摘要
Robust EEG motor imagery decoding remains limited by strong inter-individual variability, making it difficult to identify pipelines that generalize across users. We present a large-scale, standardized within-session benchmark of decoding pipelines across three public datasets: Cho2017 (52 subjects), PhysionetMI (109 subjects), and Zhou2016 (4 subjects). Using a common MOABB LeftRightImagery setting, two frequency bands (8-15 Hz and 8-30 Hz), and a broad combination of feature extraction, preprocessing, and classification steps, we analyzed 216,714 raw evaluation rows, which after structured aggregation yielded 44,928, 109,000, and 4,192 subject-level observations respectively. Covariance tangent-space projection (cov-tgsp) and Common Spatial Patterns (CSP) consistently defined the strongest methodological families, though their relative ordering was dataset-dependent. On Cho2017, the best family-level mean accuracy came from cov-tgsp in 8-30 Hz (0.712 +/- 0.140), whereas Zhou2016 favored CSP (0.832 +/- 0.121 in 8-15 Hz). These aggregate rankings concealed substantial subject-level heterogeneity: 42 distinct winning pipelines across 52 Cho2017 subjects, and 93 across 109 PhysionetMI subjects. We then used the benchmark as an empirical performance landscape for building compact portfolios of pipelines of size K. Several construction procedures were compared, including a ranking-based Top-K Mean heuristic and search-based strategies. Results were broadly consistent, with Top-K Mean giving the best trade-off. A single best global pipeline already retained 94.2% of the oracle in Cho2017 and 81.8% in PhysionetMI; at K = 12, oracle retention rose to 96.5% and 90.0%. The landscape is therefore subject-dependent, and this heterogeneity can be exploited through compact portfolios that make personalization more feasible.
