Hi Andreas and SPMers,
andreas joachim bartsch wrote:
> Dear xhosts, seeds & resoluters, particulary MGH-experts:
> Finally, I got the optseq run on my SuSE6.3-LINUX. However, I
> encountered a couple of important details to ask for:
> 1. For a simple efMRI-design specification, Matlab has properply started
> up printing out the "seed", "Eff" and "trmin". I can figure out the
> latter 2 but what is the "seed"? And what does the "Eff" [ectiveness]
> number tell me precisely, i.e. in a statistical terms for the
> feebleminded?
The Eff (efficiency) is the factor by which the variance of the noise
inherent in the experiment will be reduced by your design. Put another
way, you collect multiple observations in order to average and so reduce
the variance of your average. The efficiency is a measure of how much
variance reduction you will get.
>
> 2. Matlab terminates by telling me 4 times:
> xhost: unable to open display "iconic"
> What´s it trying to do?
optseq is calling matlab and running a matlab script. Normally, matlab
will display a "splash screen". To prevent this, I pass a "-display
iconic" flag when invoking matlab. I don't know why it's not working on
your system, but it is not really important.
>
> 3. Nevertheless, optseq generates 2 files according to my specifications
> for the output. The .dat-file looks like the trial ordering: 3 columns,
> with the second determining trial conditions and the third event
> duration in [s] for trials. The first column looks like the time axis in
> [s], right? But why is there a negative value for duration of the last
> event? And what does the log-file (i.e. tsearched, nsearched) tell me?
>
You have the interpretations of the columns correct. The duration of the
last stimulus is negative due to a bug that I've not gotten around to
fixing.
> Sorry, but I have forgotten to include another related question:
>Is there a way to use a given design to calculate optseq´s "Eff" and to
>compare it to the "optimal" suggested by optseq?
.>andreas
You can run 'seqefficiency' to determine the effiency of an already
existing paradigm.
Hope this helps
doug
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|