help needed in ICA
- hi guys,
my thesis work involves artifact correction from EEG data using
ICA. EEG data consists of 12 eeg channels and 4 reference channels
(Resp,EKG,EMG,EOG). For ICA decomposition i gave 10 sec data(1850
samples) across all 16 channels. i observed that most of the ica
components had a lot of high frequency content. for artifact
correction i visually decided the artifact components and projected
the original data back without the artifactual(supposedly)
components. The recovered data was not as clean as expected. can u
suggest me where i might be going wrong?....is the data to train the
any text or web reference will be helpful.
thanx for all the help in advance.