My name is Roey. I’m writing to you regarding the incorporation of TEDANA within fMRIprep. First of all, thank you so much for all the effort you’ve put into this!
Following issue 2542 on fmriprep’s github, it seems to me that I would have wanted to run TEDANA denoising after:
EchoN → STC → HMC → Derivative echos
Please correct me if I’m wrong, but my understanding is that currently fmriprep outputs the single echos for TEDANA after suceptibility distortion correction. Is that right?
Do you think it’s proper to apply the TEDANA denoising to these single echos, and then simply apply the transform from the original BOLD to T1w space?
Thank you so much for responding so quickly, @effigies! @effi
I followed your advice and applied the transform “boldref_to-T1w” transform.
The only thing that bothers me is this:
As far as I can tell, the single echos are before distortion correction, and the boldref_to-T1w includes the SDC. I would greatly appreciate it if you could confirm this, so I make sure I actually do what I intend to do.
I see. So there is in fact a slight worry about applying the tedana denoising approach since the SDC has already been applied.
Thank you for this clarification!
This is something that the tedana developers have discussed a fair amount. Recently, we have changed our recommendation from “run tedana before distortion correction” to “make sure you apply distortion correction consistently across echoes” (see About multi-echo fMRI — tedana 23.0.2 documentation). As far as I know, no one has done the necessary experiments to directly determine if distortion correction introduces any problems, but our belief is that the effect will be minimal, if anything.