Hi Niels,

I believe I see how that could be a problem for the current version of the toolbox (the toolbox expects to perform task-related analyses of connectivity so it will have a problem with some task conditions not being defined for a subset of subjects). In any way, for your purposes (where you are not trying to perform task-related analyses of connectivity but rather would like to simply remove any task-related effects from the BOLD timeseries) you should be able to simply import the SPM task information as first-level covariates (instead of as conditions). I have modified the conn_importspm.m script (attached) to allow you to do just that. Before performing the import step go to the matlab command line and type:
 
  global CONN_x;
  CONN_x.Setup.spmascovariate=true;

and then run the 'import from SPM' process as usual. That should import any task-related effects (as defined in SPM) as 'first-level covariates', instead of importing them as 'conditions', into the conn toolbox project information. Then fill-in manually the 'conditions' information (simply select all subjects and add a new condition named 'rest' and set 'onset' to 0 and 'duration' as inf). Last simply make sure that the newly created covariates (named 'SPM effects', for task effects, and possibly another named 'SPM covariates' or 'realignment', for subject movement effects) are present in the confounds list during the preprocessing step (and that will remove these task-related effects from the BOLD timeseries before computing any connectivity measures, implementing the process described in Fair et al.)

Let me know if you run into any issues (if this works fine I will add this as an option into the GUI and batch commands for the benefit of other users in the same or similar situation)

Best
Alfonso

Originally posted by Niels Foit: