> it will be fair enough to use the same parametric modulation for the all regressors included in contrasts (e.g. I and B),
> if we are sill interested in looking at some contrast (e.g. S-I, S-B)? Alternatively, we can use the parametric modulation
> only for a specific regressor, without considering any contrast (e.g. S-I, S-B)?
I'm not familiar with the experiment specifics, so I can't say whether it's reasonable to apply the same modulation for all the conditions. You have to make the judgment call.
In addition, when modulation is involved, the notions of S-I and S-B are not necessarily meaningful unless an index is specified. Again, you can't compare orange (intercept) to apple (slope).
> Also, as I previously mentioned, I got the error message related with the selector for the indexes (index 1 for SPMG1
> and indexes 2 and 3 for SPMG2). However, I still got the sub-briks of those contrasts, even if the error showed up.
3dDeconvolve may tend to generate something, but it is not clear to me what the results mean. Regardless, if what you specified is not meaningful, there is no point pursuing (or talking about) such a dangling end.
Gang