[VoxBo] vbview bug report

[Charan Ranganath, Ph.D.]

Dan, this is great, thanks! Regarding the overlay/GLM thing, you got 
it--I meant that if you hit the overlay button and select a .prm file, 
you go to the contrast widget and if you select a .cub, .img, or 
.whatever else file, vbview just plops it on top of the anatomical. I 
should have also said that once the kinks are worked out of VBVIEW, this 
will be a terrific addition to the voxbo suite.

A totally separate question for you/Geoff--I know that SPM deals with 
the whole autocorrelation thing with the AR(1) model. I believe (not 
sure though) that the model works by estimating the temporal 
autocorrelation of the residuals in a given GLM analysis. This would be 
useful for situations in which we don't have an empirical estimate from 
a separate dataset and don't want to use a stock function (estimated 
from the ancient GE scanners with prototype gradients from 1995). Here's 
my question, though: would we basically be doing the same thing if we 
fit a 1/f function to estimate autocorrelation from the data that we are 
trying to model? In other words, if you do a memory task and then put in 
the PS and reference function to fit a 1/f model to that, would that 
basically do the same thing as the AR(1) model?

Thanks Dan, CR


Daniel Y Kimberg wrote:
> Charan, thanks for the feedback.  I'd rather hear about these issues
> this way than find out years later that everyone's been frustrated by
> some really trivial bug (has happened a few times).  Here are some
> thoughts:
> 
>> (1) We found that, once plotted, the fitted response for any covariate
>> (or the overall fits) do not update when you change the voxel to be
>> plotted. This initially gave me quite a scare.
> 
> Will make less scary.
> 
>> (2) I would like to be able to plot the fits on the same graph as the
>> trial-averaged data. I thought I'd be able to do this by ctrl-clicking
>> the plots of the trial average and the fits. Doing this leads the fit to
>> be plotted, but the trial average is plotted as a straight line
> 
> I'll take a look at this, usually this means a scaling problem due to
> a big difference in magnitudes, and sometimes you can fix it by
> removing covariates of no interest (critically the intercept).  Other
> times not.
> 
>> (3) I could not find how to change the scale (IRF-style vs. % change vs.
>> % from zero, etc.)
> 
> Not totally available yet, I'll bump it up on the list.
> 
>> (4) I mentioned this before, but here it is again: once a GLM has been
>> loaded, the program is unable to load a new GLM. Also, it fails to
>> acknowledge changes to averages.txt after the GLM has been loaded (but
>> before the time series panel has been opened).
> 
> Still not fixed, should be soon -- 1.8.3 was an emergency release, so
> a lot of important stuff that wasn't quite ready didn't make it in.
> 
>> Other questions/comments:
>> (*) The plotting of time courses for a masked region is excruciatingly
>> slow. I'm not sure why this is the case for fairly small masks.
> 
> Yeah, I'm still playing with the code to do that efficiently.  It will
> get a lot faster at some point, but I'll see if I can make it
> moderately faster for now.
> 
>> (*) What is the use of the "interval" field in averages.txt? If your
>> events are coded in TRs, it would seem that this field should be
>> unnecessary.
> 
> Whether you're using TRs or time, you can ask it to over- or
> under-sample your time series.  The default is 1 volume, but if you
> want a different resolution, you can get it.  You also don't have to
> start your trials on TR boundaries.
> 
>> (*) When coding the onset of events in terms of TRs, should we
>> assume that the first event is "0" or "1"? (e.g., for an event
>> occurring on the 20th TR, would we code that as 19 or 20?) This
>> should be added to the documentation of the averages.txt file.
> 
> The first event is zero, the 20th TR is 19.  Will adjust help for
> averages.txt (Dongbo, you can make this change in vbcfx as well once
> I've committed my fix).
> 
>> (*) I'd like to see the "overlay" and "view GLM" options integrated if
>> this is not a major pain.
> 
> Not sure what you mean by this -- do you just want to be able to surf
> for any file to use as an overlay, and if you select a prm file it's a
> GLM, otherwise it's an overlay?
> 
>> (*) I'd like to see the return of the handy "find region" and "find
>> all" features.
> 
> You can do "find region at crosshairs," which is pretty close to the
> former.  The region stuff is still pretty messy in vbview, but I'll
> hack the "Find All Regions" button to also create a combined region
> that includes everything.
> 
>> Finally a totally separate request:
>> vbi does not seem to be able to read the origin from .img files (nifti
>> format) that are produced by SPM. I also don't believe that the correct
>> origin is imported by vb2tes. We can certainly manually set the origin,
>> but will Voxbo be able to read the origin in the future?
> 
> That's weird, but there is at least one other serious bug in how we
> process NIfTI files associated with SPM (try loading SPM5's
> templates).  I'm going to take a look at that shortly.
> 
> dan
>