[Mne_analysis] Envelope correlation

[Denis-Alexander Engemann]

        External Email - Use Caution        

One approach is to use some non-parametric stats such as Mann-Whitney-U or
Wilcoxon signed rank tests, depending on the question. You can also avoid
orthogonalization by keyword argument if you really need to satisfy the
positive definite constraint.

HTH,
Denis

On Thu 22 Aug 2019 at 13:02, Vivek Sharma <vivek.sharma1510 at gmail.com>
wrote:

>         External Email - Use Caution
>
> Hi,
> What kind of statistical test can be applied to compare among the
> different envelope correlation matrix?
> I wanted to calculate how significant is the difference between the two
> matrices and in general, these require positive definite
> Thanks
>
> On Thu, Aug 22, 2019 at 12:40 PM Alexandre Gramfort <
> alexandre.gramfort at inria.fr> wrote:
>
>>         External Email - Use Caution
>>
>> as an orthogonalization is performed to fix the leakage the matrix is
>> not positive definite
>> and a correlation matrix per se anymore.
>>
>> HTH
>> Alex
>>
>> _______________________________________________
>> Mne_analysis mailing list
>> Mne_analysis at nmr.mgh.harvard.edu
>> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/mne_analysis
>>
>
>
> --
> Vivek Sharma
> _______________________________________________
> 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/20190822/763b4d79/attachment.html