fMRIprep multiband processing problem

, ,
1

Hello Neurostars community,

I am new to fMRI preprocessing so my apologies if this is a simple question. I have processed some data using fMRIprep and would like thoughts about our output.

Here is an example of a report from a representative subject.

example
example.jpg850×1100 654 KB

Note the appearance of “bands” on some images in the ICA-AROMA section. Is this an error relating to slice time correction? We attempted to correct for this in our processing, but I am not convinced based on this output.

Any input is appreciated. Thanks!

Joseph

2

I don’t think this is a naive question, it is actually pretty hard to interpret the output of ICA-based denoising schemes.

If you are referring to components 7 and 12 (those look the most “stripy” to me), they look like noisy components you would find in a multiband dataset. If there’s a problem with slice timing correction, I’d expect to see the stripy pattern to spread over the slices.

Please note that noise components are represented in red and signal components in green color.

3

Hi @Joseph_Therriault, did we successfully reply to your question?

4

Yes! Thank you for your reply.
Joseph