Posted by Lukas Pezawas on November 19, 2002 at 11:23:58:

Hi AFNI experts!

We are conducting a mixed trial (several pure event related "block" with a baseline period in between, 2 runs of 183 volumes, TR=3000msec; the stimulus is on for 1500 msec=1/2 TR)

As far as I understood 3dDeconvolve there are 2 ways, how I can set up the contrast matrices:

1) AUC: one line with baseline columns depending of the polynomical order of the model and columns for the stimuli depending od the time lags I am using.
Since we have a rather long TR there might be only a few time points to estimate the IRF (e.g. minlag=0, maxlag=3-4, which are 4-5 time points)
E.g. If want to create a contrast of simulus A-B the matrix should look something like this: 0 0 1 1 1 1 -1 -1 -1 -1

2) One line for each time lag displaying columns with baseline parameterts
and columns displaying the contrast for each time lag separatly.
E.g: This should look like this:
0 0 1 0 0 0 -1 0 0 0
0 0 0 1 0 0 0 -1 0 0
0 0 0 0 1 0 0 0 -1 0
0 0 0 0 0 1 0 0 0 -1

I wonder, which would be from a theoretical point of view the best procedure for a trial with a long TR and a few time points as inour case?


Thanks


Lukas