Hi,
you could also take a look at the mask.img image generated by SPM analysis. If
it is not continous, you can get such kind of disconenected dot results. If
that is the case, the problem could be the thresholding or that you were using
images with very small values with non float format.
Hope this helps
Carles Falcon
University of Barcelona
Missatge de d gitelman <[log in to unmask]>:
> Dear Jie
>
> Assuming you have smoothed the data, then most likely there is some type of
> noise in your dataset. You could try plotting some of the time series or
> looking at it with tools such as tsdiffana. See this page:
> http://imaging.mrc-cbu.cam.ac.uk/imaging/DataDiagnostics
>
> Darren
>
>
> > -----Original Message-----
> > From: SPM (Statistical Parametric Mapping)
> > [mailto:[log in to unmask]] On Behalf Of J Zhuang
> > Sent: Tuesday, April 29, 2008 9:11 AM
> > To: [log in to unmask]
> > Subject: [SPM] "dot" problem of brain activation in SPM5
> >
> > Dear SPM experts,
> >
> > I met an analysis problem in the random effect of one of my
> > experimental conditions with SPM5. The activation is made up
> > of many separated tiny dots, rather than one or several big
> > blobs (see the attachment). This is the only problematic case
> > among all my experimental contrasts within the same model.
> > Here is some further information,
> >
> > The problematic contrast:
> > FWHM = 1.6 1.6 1.5 mm mm mm; 0.8 0.8 0.8 {voxels};
> > Volume: 1379328; 172416 voxels; 337791.5 resels Voxel size:
> > 2.0 2.0 2.0 mm mm mm; (resel = 0.47 voxels)
> >
> > One of the normal contrasts:
> > FWHM = 12.5 12.9 12.2 mm mm mm; 6.3 6.4 6.1 {voxels};
> > Volume: 1379328; 172416 voxels; 649.0 resels Voxel size: 2.0
> > 2.0 2.0 mm mm mm; (resel = 246.41 voxels)
> >
> >
> > I checked the fixed effect of this specific contrast for each
> > subject, but did not find any similar "dot" problem. I feel
> > puzzled with it. Could anybody be kind to give me a hint on
> > the reason and possible solution? Any help is greatly appreciated!
> >
> > Cheers,
> > Jie
> >
> >
> >
>
-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/
|