研究文章

EEGIFT:集团独立分量分析与事件相关的脑电图数据

表1

重建精度。表总结了重建精度(RA,意味着在数据集±S.E.M.)不同组的ICA模型。RA代表ICs的可变性来源占的比例平均在20个数据集,并计算分别为整个单一试验图像,周围的振幅调制在试验组件高峰延迟,平均timeseries组件,和地形。

单一的试验 高峰延迟 平均 地形
算法,Nr IC S1 S2 S3 S1 S2 S3 S1 S2 S3 S1 S2 S3

Infomax单身,20 0 9 8 ± 0 1 0 9 7 ± 0 2 0 9 8 ± 0 1 0 9 9 ± 0 1 0 9 9 ± 0 1 0 9 9 ± 0 2 1 ± 0 0 1 ± 0 0 1 ± 0 0 0 8 9 ± 0 3 0 8 9 ± 0 4 0 9 0 ± 0 3
Infomax 10 0 8 7 ± 0 6 0 7 9 ± 0 8 0 6 6 ± 0 9 0 9 2 ± 0 3 0 8 9 ± 0 3 0 7 9 ± 0 5 0 9 9 ± 0 1 0 9 4 ± 0 3 0 8 4 ± 0 4 0 8 4 ± 0 4 0 8 2 ± 0 3 0 8 1 ± 0 3
Infomax, 20 0 8 8 ± 0 4 0 7 7 ± 0 7 0 5 4 ± 0 8 0 9 1 ± 0 3 0 8 7 ± 0 3 0 6 9 ± 0 6 0 9 9 ± 0 1 0 9 6 ± 0 2 0 8 4 ± 0 5 0 8 7 ± 0 4 0 8 3 ± 0 5 0 7 4 ± 0 5
Infomax 30 0 8 7 ± 0 7 0 7 6 ± 0 7 0 5 0 ± 1 0 0 9 1 ± 0 4 0 8 6 ± 0 5 0 6 5 ± 0 5 0 9 9 ± 0 1 0 9 5 ± 0 2 0 8 6 ± 0 5 0 8 7 ± 0 4 0 8 1 ± 0 5 0 8 3 ± 0 4
Infomax 40 0 8 7 ± 0 8 0 7 5 ± 0 9 0 4 7 ± 0 9 0 9 1 ± 0 3 0 8 6 ± 0 4 0 6 2 ± 0 8 0 9 9 ± 0 2 0 9 6 ± 0 1 0 8 7 ± 0 4 0 8 8 ± 0 3 0 7 9 ± 0 4 0 8 1 ± 0 4
Infomax 50 0 8 7 ± 0 5 0 7 5 ± 0 6 0 4 4 ± 1 1 0 9 1 ± 0 5 0 8 5 ± 0 4 0 5 9 ± 0 7 0 9 9 ± 0 1 0 9 6 ± 0 1 0 8 4 ± 0 5 0 8 8 ± 0 4 0 8 1 ± 0 5 0 8 1 ± 0 4
艾丽卡,20 0 8 8 ± 0 3 0 7 3 ± 0 5 0 4 6 ± 0 9 0 9 2 ± 0 5 0 8 5 ± 0 5 0 6 2 ± 0 8 0 9 9 ± 0 1 0 8 9 ± 0 4 0 7 7 ± 0 6 0 8 6 ± 0 3 0 7 6 ± 0 6 0 7 1 ± 0 6
玉,20 0 7 9 ± 0 6 0 7 6 ± 0 7 0 4 3 ± 1 0 0 8 5 ± 0 3 0 8 7 ± 0 5 0 5 9 ± 0 9 0 9 9 ± 0 2 0 9 6 ± 0 2 0 7 9 ± 0 6 0 8 7 ± 0 5 0 7 8 ± 0 8 0 7 7 ± 0 7
fastICA, 20 0 8 6 ± 0 9 0 7 7 ± 0 8 0 5 3 ± 0 9 0 9 0 ± 0 6 0 8 7 ± 0 7 0 6 8 ± 1 0 0 9 9 ± 0 1 0 9 6 ± 0 1 0 8 2 ± 0 5 0 8 7 ± 0 4 0 8 4 ± 0 6 0 7 4 ± 0 7
SIMBEC, 20 0 8 7 ± 0 8 0 7 8 ± 0 8 0 5 0 ± 0 9 0 9 1 ± 0 6 0 8 8 ± 0 6 0 6 5 ± 0 7 0 9 9 ± 0 1 0 9 2 ± 0 3 0 7 8 ± 0 6 0 8 7 ± 0 3 0 7 9 ± 0 4 0 7 5 ± 0 5