[Eeglablist] EEG classifier for transfer learning

[Aleksandra Vuckovic]

Hi
I have 20 people in two groups performing motor imagery (of legs), group A and group B. They have different neurological conditions which are reflected in differences in ERD/ERS among these groups.

 I would like to create a classifier which would tell me whether a new person (person 21) belongs to group A or group B. This should be used for a diagnosis of a neurological condition, that is not otherwise easy detectable.

Could somebody recommend related published literature? I would like to know would be the best way to normalise data and to reduce the number of features? Which classifier should be used to prevent overfitting (there are 10 people in each group, each with 60 repetition of imagined leg movement).

Thanks,
Aleksandra




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20160607/d34bff2f/attachment.html>