Hi Chirag,
That the number of components varies per run does not really matter. You would not want to simply concatenate them, since that would force the terms that happen to go in the same column to have the same beta weight. They should probably be zero-padded across all runs, and passed using multiple -ortvec options to 3dDeconvolve.
One way to do this would be to loop over the runs using "1d_tool.py -pad_into_many_runs". See
Examples 4a and 4b from the 1d_tool.py -help output.
Does that seem reasonable?
- rick