Group classification

[Toluwani Amos]

Dear all,

I'm new to using CosmoMVPA, but i have followed the demo code stated here: (https://github.com/CoSMoMVPA/CoSMoMVPA/blob/master/doc/source/faq.rst#run-group-analysis)  for classifying between two groups. However, my tfce_map.samples shows zeros for all the voxels. Am i doing something wrong? or does this mean none of the voxels were significant?

Thanks in advance

Kind regards,

Joan

[Nick Oosterhof]

Greetings,

> On Oct 30, 2023, at 09:39, Toluwani Amos <amos...@gmail.com> wrote:
>
> I'm new to using CosmoMVPA, but i have followed the demo code stated here: (https://github.com/CoSMoMVPA/CoSMoMVPA/blob/master/doc/source/faq.rst#run-group-analysis) for classifying between two groups. However, my tfce_map.samples shows zeros for all the voxels. Am i doing something wrong? or does this mean none of the voxels were significant?

It is well possible that none of the voxels were significant after TFCE multiple comparison correction. I assume you used a group brain mask?

To verify that your analysis pipeline works as expected, you may try to add an artificial signal. For example to all participants in one group to a set of voxels chosen from an artificially defined region. If the artificial signal is made strong enough this should result in a significant cluster.

[Toluwani Amos]

Yes, i used a group brain mask.

Alright, thanks for the suggestion, i'll try this out.

Best,

Joan