Hi,
On 3 Jun 2007, at 06:54, Ping-Hong Yeh wrote:
> Dear FSLers.
>
> Our cluster output _maxc_ from randomise of comparing two groups
> difference of tbss is very asymmetric,
Do you mean the histograms from two opposite contrasts look very
different or that a single histogram is not internally symmetric?
The mapping from t to p, whilst one-to-one, is very nonlinear, so you
wouldn't expect the histograms to look similar between t and p.
Note also that t can be + or -, whereas p is from 0:1.
I would not convert to z via p, I would use ttoz, but I wouldn't
recommend this anyway - the only valid way to get cluster p-values in
TBSS is via randomise, ie using the maxc images.
Steve.
> i.e. more significantly only one
> side, though the uncorrected tstat is more bi-laterally. Any
> explanation?
>
> I tried to convert the vox_tstat (1-p image) to z-stats using ptoz
> and then
> "cluster" the z-stats image through cluster command. The group
> difference
> results was more or less like tstat image, i.e.bilaterally symmtric.
>
> Would this approach make sense for doing the cluter-base
> thresholding ?
>
> Thanks.
> PHY
------------------------------------------------------------------------
---
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
------------------------------------------------------------------------
---
|