permutation cluster analysis only shows significant results on right hemisphere

✉️ Mailing List Archive (read-only)
1

Can anyone who is familiar with permutation cluster analysis help me on
this?
I used 'permutation_cluster_1samp_test' in MNE python on my data and
only the right hemisphere showed significant results. Is it possible
that the effects on right hemisphere is so strong that the weaker
effects on the left was washed out by the permutation significant test?
Is so, can I run the permutation cluster analysis for each hemisphere
separately?

Thanks,
Xing

2

Hi Xing,

the probem is that the two hemispheres don't share any connections as
defined by the spatial connectivity matrix.
under favorable circumstances you might discover two clusters, one in each
hemisphere, which, functionally, belong togethe.
often this is not the case though.
sometimes using the `step_down_p` paremter can help which might lead to
detecting the second cluster.
separate tests for each hemisphere are possible, but I fear you will then
need to apply e.g. a bonferroni correction.
you can only test one hemisphere like that:

http://martinos.org/mne/stable/auto_examples/stats/plot_cluster_stats_spatio_temporal_repeated_measures_anova.html

HTH,
Denis

2014-09-08 20:07 GMT+02:00 Xing Tian <xt235 at nyu.edu>:

Can anyone who is familiar with permutation cluster analysis help me on
this?
I used 'permutation_cluster_1samp_test' in MNE python on my data and
only the right hemisphere showed significant results. Is it possible
that the effects on right hemisphere is so strong that the weaker
effects on the left was washed out by the permutation significant test?
Is so, can I run the permutation cluster analysis for each hemisphere
separately?

Thanks,
Xing
_______________________________________________
Mne_analysis mailing list
Mne_analysis at nmr.mgh.harvard.edu
Mne_analysis Info Page

The information in this e-mail is intended only for the person to whom it
is
addressed. If you believe this e-mail was sent to you in error and the
e-mail
contains patient information, please contact the Partners Compliance
HelpLine at
MyComplianceReport.com: Compliance and Ethics Reporting . If the e-mail was sent to you in
error
but does not contain patient information, please contact the sender and
properly
dispose of the e-mail.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20140909/dc0d1f39/attachment.html

3

hi,

as denis suggested I would give a try to step_down_p.
Non-parametric cluster stats tend to be dominated by strongest effect
which you can sequentially exclude using the so called "step-down" procedure.

Alex