本文详细介绍EEG信号预处理的流程，包括高密度EEG数据收集、导入Matlab、事件标记与通道位置导入、重新参考与降采样、高通滤波、原始数据检查、坏通道剔除、大工件时间点拒绝及独立成分分析。揭示了良好ICA分解的秘诀，如避免非典型大型工件、确保足够数据量、高通滤波去除缓慢漂移等。
翻译水平有限，写的英文的~EEG信号预处理基础知识
预处理流程（pre-processing）
1.collect high-density EEG data
2.import into matlab
3.import event markers and channel locations
4.re-reference/down-sample(if nessary)
5.high pass filter(0.5-1 Hz)
6.examine raw data
7.reject bad channels
8.reject large artifact time points
9.RUN ICA
## Secrets to a good ICA decomposition
- garbage in … garbage out（it’s not magic）
- remove large,non-stereotyped artifacts
- enough data?(based mostly on time,not frames)
- ~30 min of data for 69-70 channels, ~60 min for > 200 channels
- high-pass filter to remove slow drifts
- low-pass/ notch filters usually unnecessary
- remove bad channels
- data must be in double precision (not single)
- Tools–》Auto-detection of noisy channels
- interpolate bad channels
源自EEGlab官网
扫一扫
539
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
