Dear Alfonso and other CONN users,

Hopefully a simple question - where are details regarding one's chosen ART settings recorded in the CONN_x file?

It appears some files have a CONN_x.SetupPreproc.log cell array where this information is stored; but I seem to also have old projects where the "log" is not present - in these cases is there another location where this information may be available?


Many thanks

Rob

Dear Rob,

Sorry about that but you are right that in old-enough versions of CONN that information was not being stored anywhere. A few options below:

1) if you run your preprocessing steps using parallelization options (either on your own computer on the background or in a distributed cluster), then you find that information in the node_*.mat files within the corresponding conn_*.qlog/#datestring# subdirectory

2) if your version creates the second-level covariates "QC_MeanMotion" and "QC_MeanGSchange", then you can also get that info from the description field of these second-level covariates, even if the log was not yet being created.

3) if your version creates a first-level covariate named 'QC_timeseries' containing the raw (unthresholded) motion and GSchange timeseries, you could then 'reverse-engineer' which threshold was used by re-creating the thresholded scrubbing covariate from the QC_timeseries covariate using different threshold values and seeing which one produces the same results as your original scrubbing covariate (to do that you may use the syntax "conn_convertl12l1covariate('scrubbing',[3 0.5],1,'QC_timeseries','new_scrubbing')" using the attached file in order to generate the new "new_scrubbing" covariate using the [3 0.5] threshold values)

If none of that works in your case, let me know which version you have and I will see if there are any other simple alternatives.

Hope this helps
Alfonso


Originally posted by Rob McCutcheon:

Dear Alfonso,

Many thanks for taking the time to reply

"2) if your version creates the second-level covariates "QC_MeanMotion" and "QC_MeanGSchange", then you can also get that info from the description field of these second-level covariates, even if the log was not yet being created."

This is likely the best option as the analysis was ran without parralleiisation and I only have 'realignment' and 'scrubbing' as first level covariates (no QC_timeseries)

I do have the second level covariate 'QA_MeanMotion", however in the Description field on Setup->Covariates it simply says "CONN Quality Assurance: Average inter-scan motion observed"

Am I looking in the right place? I am using version 17b

Best Wishes

Rob