[Mne_analysis] Linear regression on Time Frequency data
Alexandre Gramfort
alexandre.gramfort at inria.fr
Mon Jul 13 08:49:54 EDT 2020
External Email - Use Caution
hi Sebastian,
the only thing I can suggest is that if you think the signal is in the
time-frequency
power than you should not process electrodes separately but use dedicated
spatial filtering techniques that can boost SNR (See CSP, SpOC in the
MNE-Python
documentation).
HTH
Alex
On Tue, Jul 7, 2020 at 11:33 AM S.P.H. Speer <speer at rsm.nl> wrote:
> External Email - Use Caution
>
> Hi Alex,
>
>
>
> I create subject level contrasts (conflict versus no conflict) and then want to run a group-level analysis where I want to use a continuous variable, which is an individual difference measure (number of trials on which the participant cheated), as a predictor in a linear regression with t-values from the within subject contrasts (conflict versus no conflict) as dependent variable. Stated differently, I want to see whether frequency of cheating across individual can explain differences how conflict versus no conflict trials are modulated in the brain.
>
> I’ve got the within-subject contrasts for each subject TF-maps one for each channel and am now wondering which function I can use to run this group-level analysis?
>
>
>
> Thanks for your help,
>
>
>
> Sebastian
>
>
>
>
>
>
>
> hi Sebastian,
>
>
>
> can you step back and explain the overall aim of the procedure?
>
> what would you like to show with your data? is it group level?
>
> a contrast?
>
>
>
> Alex
>
>
>
>
>
> On Mon, Jun 29, 2020 at 6:14 PM S.P.H. Speer <speer at rsm.nl <https://mail.nmr.mgh.harvard.edu/mailman/listinfo/mne_analysis>> wrote:
>
> >
>
> >* External Email - Use Caution*
>
> >
>
> >* Dear MNE-team,*
>
> >
>
> >
>
> >
>
> >* I’m trying to run a linear regression analysis on Time-Frequency data. I’ve got the average TFR for each for each subject and want to regress and individual difference measure on the time frequency data for a specific channel. The data has the same dimensions as for running linear regression on epochs but instead of channels I’m using frequencies. Unfortunately the linear_regression function provided by MNE does not allow TFR data as input. Is there a workaround for this problem? I would also like to run cluster-based permutation tests on the results of the linear regression to estimate significance. Would that be feasible?*
>
> >
>
> >
>
> >
>
> >* Thank you very much for your help,*
>
> >
>
> >
>
> >
>
> >* Sebastian*
>
> >
>
> >* _______________________________________________*
>
> >* Mne_analysis mailing list*
>
> > *Mne_analysis at nmr.mgh.harvard.edu <https://mail.nmr.mgh.harvard.edu/mailman/listinfo/mne_analysis>*
>
> > *https://mail.nmr.mgh.harvard.edu/mailman/listinfo/mne_analysis <https://mail.nmr.mgh.harvard.edu/mailman/listinfo/mne_analysis>*
>
>
> _______________________________________________
> Mne_analysis mailing list
> Mne_analysis at nmr.mgh.harvard.edu
> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/mne_analysis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20200713/5e937252/attachment.html
More information about the Mne_analysis
mailing list