label_sign_flip

Mailing List Archive (read-only)
1

Dear MNE-python users,

There's a new function in label.py called label_sign_flip. It is used in
one of the examples:

# compute sign flip to avoid signal cancelation when averaging signed values
flip = mne.label_sign_flip(label, inverse_operator['src'])

Could someone explain in a bit more detail what this function is designed
to do and when it should be used?

Thanks.
Tal
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20120604/2668b3d6/attachment.html

2

hello Tal,

mne.label_sign_flip is a Python implementation of one of the MNE approach
to compute a ?label mean" i.e. get one time series of current
estimates per label.
If you average signed currents in a label you're
likely to destroy your signal by cancelation effects. The mne.label_sign_flip
command returns a sign convention that should be applied to your signed
estimates before averaging.

see:

http://martinos.org/mne/auto_examples/inverse/plot_compute_mne_inverse_epochs_in_label.html

let me know if it's unclear.

Alex