Hello Andreja,
Can you upgrade to 5.0.8, as this is an issue which has hopefully been fixed.
Kind Regards
Matthew
> Hi Matthew,
>
> No, we are using FSL 5.0.7.
>
> Many thanks
> Anja
> ________________________________________
> From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Matthew Webster [[log in to unmask]]
> Sent: 16 December 2014 16:48
> To: [log in to unmask]
> Subject: Re: [FSL] FEAT: Errors at the first level
>
> Hello Andreja,
> Are you using FSL 5.0.8?
>
> Kind Regards
> Matthew
>> Dear fsl developers,
>>
>> I am writing because I am encountering the fatal error problem again when using FEAT to carry out the first level analysis (please see below for the email correspondence concerning this issue).
>>
>> As in my previous analysis, I get the error at the end of the preprocessing and the fatal error at the end of stats. I emailed you my raw data back in Septmeber and you advised me that these errors are benign and that I can carry out the analysis despite fatal errors, as I've got activation maps in post-stats which looked reasonable.
>>
>> I've tried running the same analysis on a new subject today: I've got the same fatal errors as before, however, this time, the analysis did not complete (the analysis stopped after completing the stats step). I also get a new error at very top of the progress report saying:
>>
>> mkdir .files;cp /usr/local/fsl-5/doc/fsl.css .files;ln -s /usr/local/fsl-5/doc/images .files/images
>>
>> /usr/local/fsl-5/bin/fsl_sub -T 10 -l logs -N feat0_init /usr/local/fsl-5/bin/feat /studies/2009/BBCEO/converted/BrainBank_0073/RUN1_FULL_SAMPLE_GR.feat/design.fsf -D /studies/2009/BBCEO/converted/BrainBank_0073/RUN1_FULL_SAMPLE_GR.feat -I 1 -init
>> 15398
>>
>> /usr/local/fsl-5/bin/fsl_sub -T 79 -l logs -N feat2_pre -j 15398 /usr/local/fsl-5/bin/feat /studies/2009/BBCEO/converted/BrainBank_0073/RUN1_FULL_SAMPLE_GR.feat/design.fsf -D /studies/2009/BBCEO/converted/BrainBank_0073/RUN1_FULL_SAMPLE_GR.feat -I 1 -prestats
>> 16205
>>
>> /usr/local/fsl-5/bin/fsl_sub -T 640 -l logs -N feat3_film -j 16205 /usr/local/fsl-5/bin/feat /studies/2009/BBCEO/converted/BrainBank_0073/RUN1_FULL_SAMPLE_GR.feat/design.fsf -D /studies/2009/BBCEO/converted/BrainBank_0073/RUN1_FULL_SAMPLE_GR.feat -I 1 -stats
>>
>> FATAL ERROR ENCOUNTERED
>>
>> Could you tell me what is the issue and whether it can be quickly resolved?
>>
>> Prompt reply is highly appreciated.
>>
>> Thanks very much
>> Anja
>>
>>
>> ________________________________________
>> From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Matthew Webster [[log in to unmask]]
>> Sent: 22 September 2014 11:30
>> To: [log in to unmask]
>> Subject: Re: [FSL] FEAT: Errors at the first level
>>
>> Hello Andreja,
>> Benign means that in this case the error can safely be ignored. I would advise you to continue with your analysis.
>>
>> Kind Regards
>> Matthew
>>
>>> Dear Matthew
>>>
>>> Thanks for a reply. Could you please clarify what does 'benign' error mean? Does this mean that the errors do not affect the analysis outputs? Would you advise me to carry on with the analysis regardless of the errors in the log? There are three errors in the log - are all of these equally 'benign'? Sorry for being a pain but I have a deadline to meet and I would need to know what does it mean to run the analysis on 20 subjects with a fatal error in the log.
>>>
>>> Thanks very much
>>> Anja
>>> ________________________________________
>>> From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Matthew Webster [[log in to unmask]]
>>> Sent: 19 September 2014 10:16
>>> To: [log in to unmask]
>>> Subject: Re: [FSL] FEAT: Errors at the first level
>>>
>>> Hi Andreja,
>>> This is a "benign" FATAL error which arises from subtle details in how FEAT handles a zero-EV, the error reporting code should handle this more gracefully in the next patch.
>>>
>>> Kind Regards
>>> Matthew
>>>> Dear Matthew
>>>>
>>>> It's been a week since I uploaded the analysis output onto the fsl website. Could you please tell me whether there are any updates on this?
>>>>
>>>> Many thanks
>>>> Anja
>>>> ________________________________________
>>>> From: Andreja Varjacic
>>>> Sent: 12 September 2014 16:42
>>>> To: FSL - FMRIB's Software Library
>>>> Subject: RE: [FSL] FEAT: Errors at the first level
>>>>
>>>> Hi
>>>>
>>>> I've uploaded a tar of the whole folder "RUN1_FULL_SAMPLE_GR.tar.gz" onto the system.
>>>>
>>>> Many thanks
>>>> Anja
>>>> ________________________________________
>>>> From: FSL - FMRIB's Software Library [[log in to unmask]] on behalf of Matthew Webster [[log in to unmask]]
>>>> Sent: 12 September 2014 15:58
>>>> To: [log in to unmask]
>>>> Subject: Re: [FSL] FEAT: Errors at the first level
>>>>
>>>> Hello,
>>>> Thank you for the upload - it looks like I will need a tar of the whole folder though.
>>>>
>>>> Kind Regards
>>>> Matthew
>>>>> Dear FSL developers
>>>>>
>>>>> I am writing because I am experiencing a number of errors when running first level analysis using FEAT.
>>>>>
>>>>> Looking at the log, there appears to be an error at the end of the preprocessing, a 'fatal error' at the end of the stats stage and a warning about smoothing in the poststats. This is what the log reads:
>>>>>
>>>>> PREPROCESSING STAGE 2:
>>>>>
>>>>> [...]
>>>>>
>>>>> /usr/local/fsl-5/bin/fslmaths prefiltered_func_data_tempfilt filtered_func_data
>>>>>
>>>>> /usr/local/fsl-5/bin/fslmaths filtered_func_data -Tmean mean_func
>>>>>
>>>>> /bin/rm -rf prefiltered_func_data*
>>>>> Completed
>>>>> while executing
>>>>> "if { [ catch {
>>>>>
>>>>> for { set argindex 1 } { $argindex < $argc } { incr argindex 1 } {
>>>>> switch -- [ lindex $argv $argindex ] {
>>>>>
>>>>> -I {
>>>>> incr arginde..."
>>>>> (file "/usr/local/fsl-5/bin/feat" line 310)
>>>>> -stats
>>>>> Completed
>>>>>
>>>>> STATS:
>>>>>
>>>>> /usr/local/fsl-5/bin/film_gls --in=filtered_func_data --rn=stats --pd=design.mat --thr=1000.0 --sa --ms=5 --con=design.con
>>>>> Log directory is: stats
>>>>> paradigm.getDesignMatrix().Nrows()=491
>>>>> paradigm.getDesignMatrix().Ncols()=27
>>>>> sizeTS=491
>>>>> numTS=39675
>>>>> Calculating residuals...
>>>>> Completed
>>>>> Estimating residual autocorrelation...
>>>>> Calculating raw AutoCorrs... Completed
>>>>> mode = 9966.43
>>>>> sig = 1261
>>>>> Spatially smoothing auto corr estimates
>>>>> .....................
>>>>> Completed
>>>>> Tukey M = 22
>>>>> Tukey estimates... Completed
>>>>> Completed
>>>>> Prewhitening and Computing PEs...
>>>>> Percentage done:
>>>>> 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,Completed
>>>>> Saving results...
>>>>> Completed
>>>>>
>>>>> FATAL ERROR ENCOUNTERED
>>>>>
>>>>> (then it carries on to the post stats)
>>>>>
>>>>> POST-STATS:
>>>>>
>>>>> /usr/local/fsl-5/bin/fslmaths stats/zstat1 -mas mask thresh_zstat1
>>>>>
>>>>> /usr/local/fsl-5/bin/smoothest -m mask -z stats/zstat1 > stats/zstat1.smoothness
>>>>> WARNING: Extreme smoothness detected in X - possibly biased global estimate.
>>>>> WARNING: Extreme smoothness detected in Y - possibly biased global estimate.
>>>>> WARNING: Extreme smoothness detected in Z - possibly biased global estimate.
>>>>>
>>>>>
>>>>> The confusing part is that feat carries on with the next stage of the analysis as if the errors were not detected at all. When the analysis is completed, I get the motion/stats/post stats/registration outputs I would expect in the error-free analysis. Technical staff from our department tried to track the errors but to no avail: it looks like feat reports errors which, when inspected, do not exist.
>>>>>
>>>>> I have ran this analysis protocol on over 50 subjects in the past two years and I had no issues like this at all. I started getting these errors after fsl update in the mid July. This may have caused the errors, but nobody understands how.
>>>>>
>>>>> Could you please help? Would you advise me carrying on with the analysis despite the errors (as they *appear* not to affect the data very much) or can these errors be tracked and resolved?
>>>>>
>>>>> Thanks very much,
>>>>> Anja
|