Hello Michael,
Your all_FA_skeletonised.nii.gz consists of the same image repeated 12 times - could there have been an error in merging the individual files? By default randomise 2.9 will detect and remove voxels with a constant timeseries from the data - which in this case is all the voxels! You can disable this behaviour with --norcmask, but I think you will want to check how the all_FA_skeletonised image was generated first.
Hope this helps,
Matthew
> Matthew,
>
> Okay, great. I uploaded the data. The file name is:
>
> upload_michael_fsl.tar.gz
>
> The 6-digit reference number is: 294333
>
> Please let me know, if you need any additional information.
>
> Thanks,
>
> Michael
>
> On Mon, 9 Jan 2012 15:50:54 +0000, Matthew Webster <[log in to unmask]> wrote:
>
>> Hello Michael,
>> If you want to upload your input files to:
>>
>> http://www.fmrib.ox.ac.uk/cgi-bin/upload.cgi
>>
>> and let me know the 6-digit reference number, I will check to see if this is a bug introduced into randomise 2.9.
>>
>> Many Regards
>>
>> Matthew
>>> One more thing, regarding this issue,
>>>
>>> I'm currently running fsl 4.1 (build 419) and randomise v2.9. I installed my fsl package via Debian.
>>>
>>> This is where I am getting the error message.
>>>
>>> Just on a hunch, i ran the same command line on our old server that had FSL 4.1 (build 416) and randomise v2.6, and I received no error message. The analysis actually worked.
>>>
>>> So,I'm not so sure what is going on. Is there a bug in v2.9, possibly?
>>>
>>> My older version of randomise is installed on Redhat 2.6.9.
>>>
>>> I've installed build 419 of FSL via the Debian package where we are running Ubuntu. Could that be an issue?
>>>
>>> Michael
>>>
>>> On Mon, 9 Jan 2012 14:13:29 +0000, Michael Keaser <[log in to unmask]> wrote:
>>>
>>>> Hello,
>>>>
>>>> Yes, I double checked in FSLview (and also in AFNI viewer) that both sets of masks overlay the data.
>>>>
>>>> I tried gunzipping the nii files for both the masks and the all_skelentonized data see if that was the issue, and still the same error occurs.
>>>>
>>>> Michael
>>>>
>>>>
>>>> On Mon, 9 Jan 2012 13:52:36 +0000, Matthew Webster <[log in to unmask]> wrote:
>>>>
>>>>> Hi,
>>>>> Can you double check that the mask overlaps the data ( e.g. just quickly check them in FSLview? )
>>>>>
>>>>> Many Regards
>>>>>
>>>>> Matthew
>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> Thanks for your response.
>>>>>>
>>>>>> Unfortunately, I've tried using both mask optoins mean_FA_mask and mean_FA_skeleton.nii.gz, and I still get the same error message.
>>>>>>
>>>>>> Anybody else having the same issue or is there another step that I am missing?
>>>>>>
>>>>>> Michael
>>>>>> ________________________________________
>>>>>> From: Haitao Ge
>>>>>> Sent: Saturday, January 07, 2012 8:47 AM
>>>>>> To: [log in to unmask]
>>>>>> Subject: Re: [FSL] TBSS exception
>>>>>>
>>>>>> Hi Michael
>>>>>>
>>>>>> I think you have input the wrong mask. You have input the mean FA skeleton
>>>>>> instead of mean FA mask.
>>>>>>
>>>>>> Haitao
>>>>>>
>>>>>> Center for Sectional and Imaging Anatomy
>>>>>> Shandong University School of Medicine
>>>>>> 44#, West Wenhua Rd.
>>>>>> Jinan 250012, Shandong
>>>>>> P.R.China
>>>>>>
>>>>>> -----原始邮件-----
>>>>>> From: Michael Keaser
>>>>>> Sent: Saturday, January 07, 2012 5:44 AM
>>>>>> To: [log in to unmask]
>>>>>> Subject: Re: [FSL] TBSS exception
>>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> I too have had the same problem with the error mask is blank.
>>>>>>
>>>>>> Here is my command and the output from it:
>>>>>>
>>>>>> randomise -i all_FA_skeletonised.nii.gz -o pairedT -d design-glm.mat -t
>>>>>> design-glm.con -f design-glm.fts -m mean_FA_skeleton.nii.gz -e
>>>>>> design-glm.grp -T
>>>>>> randomise options: -i all_FA_skeletonised.nii.gz -o pairedT -d
>>>>>> design-glm.mat -t design-glm.con -f design-glm.fts -m
>>>>>> mean_FA_skeleton.nii.gz -e design-glm.grp -T
>>>>>> Loading Data: ERROR: Program failed
>>>>>>
>>>>>> An exception has been thrown
>>>>>> Data mask is blank.
>>>>>>
>>>>>>
>>>>>> I'm not sure what is going on. Anyone have any ideas? Your help would
>>>>>> greatly be appreciated.
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>> Michael
>>>>>>
>>>
>
|