Johanna
Are you saying that you want to use skin conductance as a regression variable? That is, to try to account for variability in your fMRI time series that may be due to changes in skin conductance? If so, the skin conductance data would enter your analysis as a .1D file and be treated as a stimulus by incrementing you -num_stimts entry and adding an additional -stim_file entry.
If your temporal resolution is different than your fMRI data (and it appears that it is) you will have to resample your conductance data so that it has the same sample rate as your fMRI data.
KJ