Try it with the correct contrast for a paired t-test. If this shows any
difference, it will indicate that the initial affine registration done prior
to the unified segmentation (but built into what goes behind the button
press) has gone wrong.
Best regards,
-John
On Tuesday 18 December 2007 14:03, Simone Reinders wrote:
> Dear list,
>
> After reading this email:
> http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind05&L=SPM&P=R466366&I=-3
> and Good et al. (NeuroImage 14, 2001, p22, 2nd column), our complete
> data set of 202 SPGR scans was reoriented (via the Display button) by
> one person. About 10-15% of the
> scans were in the range specified in the email above. After this the
> reoriented data was segmented and smoothed with 8mm.
>
> To see the effect of reorientation we performed a paired t-test pre-
> vs postreorientation of the SMWC1, SMWC2 and SMWC3 separately, with
> the settings listed below.
>
> The results are in the PDF files attached. There seem to be very
> strong differences between the pre- and postreorientation data and we
> do not know how to explain them or whether or not we should worry
> about such effects.
>
> Any suggestions are greatly appreciated.
>
> Thank you,
>
> Simone Reinders and Andy Simmons.
>
>
> King's College London
> Institute of Psychiatry (IoP)
> London, UK.
>
> - independency setting as 'no'
> - variance 'unequal'
> - grand mean scaling as 'no'
> - threshold masking: 'none'
> - impliciet mask 'yes'
> - expliciet mask '<none>'
> - global calculation 'omit'
> - global normalisation
> - overall grand mean scaling 'no'
> - normalisation 'none'
|