I would not agonize about getting the maximum possible efficiency. When you go to analyze your data, there will be differences that you cannot control (eg, including motion correction regressors or temporal whitening), and these will change efficiency. The only thing I would change to the command is to add --focb 100
doug
On 6/16/13 10:53 AM, fMRI wrote:
Dear Prof Douglas,
I asked this before but I would like to understand what I am doing and also need your recommendation.
I have three conditions : squeezes at three different levels. My aim is that I want to use spm to enter them as a one condition and then enter the parametric response as a covariate regressor. This will allow me to check the effect of using high low and mid forces using for example the first order effects.
I am trying to find the best event design using optseq. I do not mind to make the stimuli close or far away from each other.
What I have is tr 2.5 and ntp 174 so the length of the scanner is 435s.
I tried different input:
Optseq --ntp 174 --tr 2.5 --psdwin 0 20 2.5 --ev 1 3 20 --evc 2 3 20 --evc 3 3 20 --nkeep 1 --tnullmin 2 --nullmax6
The question that I have what are the inputs that I should change in order to get the best design that answer my question? I mean I thought that I should focus in tge nullmin and max. So in this. Case does increasing the null max make the design better. Also in term of the contrast should I look for it as a general term without including any contrast since I will enter them as one condition or should I see the effects of rest [1 1 1]
Any tip or advice that I should look at and change?
Lastly, if I would like to compare different designs, should I compare just the eff result ?
Thanks
AS