Hi,
Just to add to this - the changes that are seen due to version changes and OS differences are typically small. However, viewing changes in thresholded maps can sometimes over-emphasize such differences. It is therefore better to assess the changes to the unthresholded stats. Unless you are seeing large changes in the unthresholded statistics then there shouldn't be anything to worry about and your group statistics are unlikely to change much.
All the best,
Mark
On 24 Oct 2014, at 16:07, Matthew Webster <[log in to unmask]> wrote:
> Hello Kim,
> Each update to FSL will contain multiple fixes, modifications and improvements to our software, which will almost certainly result in changes to calculated statistical parameters etc. In general we recommend using the latest version of FSL for all analyses, with the potential exception of a longitudinal study where it might be preferable to use the same FSL version for the duration of the study.
>
> Kind Regards
> Matthew
>> Dear FSL experts,
>>
>> I have a problem with FEAT, first level analysis. With the same parameter and the same image, i got the different results from different versions of FSL (5.0.2 in Vmware centos and 5.0.7 in Mac). Could you please help me to explain it and which one is correct?
>>
>> Thanks a lots,
>>
>> Kim
|