Hello
I have some questions to ask.
-First is there a good montage for the SEED V dataset.
-Should we keep the channels more than 5 std dev for ICA
-Should I drop channels with no coordinates before ICA (the cannaboid channels CB1,CB2, eog channels VEO, HEO) if I do not find a montage that has them too.
-Should I use lower number of components than tha of the channels in ICA (I think I do to avoid overfitting not sure yet)
-How accurate the labaling on the kind of brain activity provides is in general ?
I follow the specific pipeline till now and the results I do not like much
-filter 1-50
-reference avarage
-ICA for 60 component
-dropped the ones that were flagged not brain (I need brain for emotion recognition)
After that I still see a lot of noise and in some places even more noise than before. Thats why I ask
Thank you
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS