Many thanks for the prompt reply, Rick. It seems to be working fine now.

However, in relation to setting up the regression, I still have some questions about BLOCK and GAM. I've read the help output for 3dDeconvolve for these options, but I admit I don't entirely understand them. I haven't had too much luck understanding other sources online either.

I modelled the task by:
1. Removing the first breathhold and 17TRs from the input file (rather than the 11TRs I had mentioned previously)

2. Entered the adjusted stim_times to accommodate the removed breathhold and TRs (mapped the task out in Excel to check, but essentially removed the first breathhold onset time and subtracted the remaining times by 68s)

3. Added to each breathhold onset the mean amount of time needed on average for the BOLD response to peak after each breathhold for that subject, as calculated in MATLAB

4. Used a BLOCK (15,1) HRF, given that each breathhold is 15s in duration

When I read the help file, it says there are two variants of BLOCK (4 and 5), which have different delays. How important is the distinction between these two functions, and is it appropriate for me to add my individualized mean delay values to each breathhold onset if the BLOCK function is applying some kind of delay?

For the GAM function, do I simply use GAM (15,1)? How does it differ from BLOCK? As with BLOCK, I don't entirely understand the help file's description of the GAM option.

Thanks,
Alain