[VoxBo] timeseries and jlextract

[Daniel Y Kimberg]

Ling Wong wrote:
> 1) I'm trying to get the timeseries for a single subject time-locked
> to the first stimulus of a trial.  When I try to view this in either
> vbview or Display, it gives me a timeseries with many timepoints --
> the same number of timepoints I have in the entire experiment.  My
> understanding was that getting a timeseries using a certain contrast
> (e.g. first stimulus is weighted +1), it averages the beta values
> that follow each presentation of that stimulus.  If that's the case,
> then why does the timeseries have as many timepoints as the entire
> experiment?

Ling, hi.  If you want trial averaging in vbview (or using vbxts at
the command line), you need to specify how you want the data averaged
in your averages.txt file.  When you run a glm, the file is created in
the GLM directory with enough comments to get you started.  Then in
vbview (or vbxts) you can just select the averaging scheme you want to
use.  I believe Display (the IDL program) will sometimes figure this
out for you, but I don't remember the details.

>From the looks of things, ksh isn't installed wherever you tried it.
If that was at the CfN, you can probably talk Chris into it.

dan