Hello,
I am trying to use the eye blink correction script included in Laurence's remove_eyeblink_topography_walkthrough.m, but I have some issues and I would very much appreciate any help since I am new to this processes.
I was able to execute the script on both your "demo" dataset and my own datasets, but the results on the latter seem to have some issues. More specifically:
1. By comparing the corrected raw signals (just after step 11, see attached Laurence1_25_after.png) and the initial ones (see attached Laurence1_25_before.png), there seems to be a change in the polarity of the first 10 EEG electrodes only, i.e. frontal EEG channels, in segments not adjacent to EOG blink artifacts. Why does this happen? Is it normal?
2. I adjusted the script to run it on my dataset with continuous recording of 60 eeg channels. It seems to detect eye blinks quite well, but the correction step (11) seems to produce distortions, like amplitude shifts on segments of certain channels and/or high frequency noise. Do you have any idea why these distortions occur? (I can send some figures or share the dataset itself if necessary)
Again any help would be very much appreciated,
Dimitra Pseftogianni
Phd candidate
Mathematics/Statistics Graduate
|