tf_mixed_norm group solution?

Mailing List Archive (read-only)
1

Hello All,

Is there a way to force tf_mixed_norm in the python package to generate a single set of localizations which can be applied to 1) multiple subjects and, 2) multiple experimental conditions within the same subject?

With regard to multiple subjects, since each subject would require an individual forward solution, even if they were being morphed to a common space in the end, this does not currently seem possible?

A single, average forward solution could be created that spanned conditions within a single subject, which could then be applied to the concatenated evoked averages for the individual conditions. The time courses of the single set of localizations could then be re-segmented into the separate conditions. This would require some scheme to minimize the artifact generated by the discontinuities, but it could be done?

Thanks for everyone's help,

-Per Lysne

University of New Mexico

lysne at unm.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20140311/95bd3e41/attachment.html

2

hi Per,

Is there a way to force tf_mixed_norm in the python package to generate a
single set of localizations which can be applied to 1) multiple subjects
and, 2) multiple experimental conditions within the same subject?

sadly no.

It has never been written although it is "just" an engineering
problem and not a "science" problem at least for the multi condition case :slight_smile:

With regard to multiple subjects, since each subject would require an
individual forward solution, even if they were being morphed to a common
space in the end, this does not currently seem possible?

indeed you could use a morphed source space but you would still
have to pass each forward solutions unless you find a way to avoid it
by some clever preprocessing. I don't know how.

A single, average forward solution could be created that spanned conditions
within a single subject, which could then be applied to the concatenated
evoked averages for the individual conditions. The time courses of the
single set of localizations could then be re-segmented into the separate
conditions. This would require some scheme to minimize the artifact
generated by the discontinuities, but it could be done?

yes that sounds really reasonable. Let us know if you need
help to look into this.

Thanks for everyone's help,

no pb

Best,
A