Dear SPMers,
I recently specified a model and then tried to estimate it, but
ran into problems. Essentially, during model estimation the progress bar
stayed at 0% for 12 hours, after which I quit the process. I've entered
the parameters I selected during model specification and estimation
below. If anyone sees what I did wrong, I'd greatly appreciate some
advice on what to change. Also, if I do need a small change, is there any
way to alter the SPMcfg.mat or SPM_fMRIDesMtx.mat file(s) that have already
been created to avoid manually typing all the parameters into the GUI
interface again (it took over an hour the first time).
Thanks,
Daniel Weissman
For the model specification, I entered the following parameters:
Interscan interval 2
Scans per session 176*ones(1,74) --is this the problem?
SPM accepted it and
understood that I wanted
74 sessions.
Are conditions replicated? No.
Then, for all 74 sessions, I entered:
# of conditions 2
name for condition 1 GC
name for condition 2 GI
Stochastic Design No
SOA Variable
Vector of onsets -GC a vector containing the onset times
for session x
Variable Durations No.
Vector of onsets -GI a vector containing the onset times
for session x
Variable Durations No.
Are these trials Events
Select basis set hrf w/time derivative
interactions among trials (Volterra) No.
User specified regressors 0
For model estimation, I
1) selected the snaV*.img files corresponding to the 74 sessions of my
study in the order the sessions were specified during model specification.
2) remove Global effects - scale
3) High-pass filter - accepted defaults
4) Low-pass filter - Hrf
5) Model Intrinsic correlations? - none
6) Setup trial-specific F-contrasts? - yes
7) Estimate - now
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|