I have taken three images per subject, with voxel values normally distributed. I then combined them by taking the magnitude across images. The result is an image with a square root of Chi-squared distribution.
I have this for ten subjects and would like to know which voxels have values consistently (across subjects) large values.
I would like to do a one group analysis but with these non-normally distributed images.
Thank you,
Jason.
----- Original Message ----
From: Steve Smith <[log in to unmask]>
To: [log in to unmask]
Sent: Thursday, July 12, 2007 3:25:57 PM
Subject: Re: [FSL] Randomise on a group of Non-normal data
Hi,
Sign flipping generates the null distribution for there being a non-
zero mean effect, which doesn't match what you have I suspect. What
do you mean by "significant" - what is your model?
Cheers, Steve.
On 12 Jul 2007, at 17:06, Jason Steffener wrote:
> Hello all,
> I have a single group of images, one per subject, where the voxel
> values are distributed as square root of Chi-squared
> distributions. I would like to perform a group level analysis on
> this data to find those voxels that are consistently significant
> across the group. I am wondering if the randomise program is
> applicable. I know that for a single group of subjects the program
> flips signs. So would this applicable for the distributions I have?
> If not does anyone have any other suggestions.
>
> Thank you for any advice.
>
> Jason.
------------------------------------------------------------------------
---
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
------------------------------------------------------------------------
---
|