--combine-runs to work with reho and alff

[CogBrainHealthLab]

Currently when the --combine-runs y flag is specified, it outputs a single fMRI volume, however it does not seem to have the same effect for the *alff_boldmap and *reho_boldmap files. Ideally, I would like to have a single reho or alff volume that is concatenated across runs.

[tsalo]

Could you clarify what you mean by ReHo or ALFF being concatenated across runs? Do you mean:

• A 4D file with one volume for each run
• An unweighted average across runs
• A weighted average across runs, with the weighting based on something like number of volumes in the original runs (before or after censoring?)
• Concatenating the time series and recalculating ALFF and ReHo

[CogBrainHealthLab]

Definitely not the first option.

would "Concatenating the time series and recalculating ALFF and ReHo" be the most advisable method to combine the ALFF and ReHo volumes from multiple runs into a single volume?

[tsalo]

I don't know the literature on combining ALFF and ReHo across runs, so I don't know what folks generally do. To be honest, I think that people should be taking the individual runs' data and using inter-run variability as a measure of subject-level variability instead of concatenating or simply averaging data, but that adds an extra level of complexity to the model.

[mattcieslak]

I think ALFF in particular may not be appropriate for concatenated runs. ALFF is frequency-based, and the discontinuity where data is concatenated may not be ok.

[CogBrainHealthLab]

what about reho?
Because if its alright to extract the parcellated FC matrix from the concatenated runs, it should make sense to do the same for reho?

[tsalo]

It seems okay as far as I can tell, but I don't want to add it without something in the literature to reference, since I don't know the math well enough to say ReHo on concatenated data is valid.