Le 23/11/2012 18:21, Jesper Andersson a écrit :
> Dear Romain,
>
>> I am tring to unwarp a double dti acquisition (AP and PA). If I extract the B0 and do the topup comand it works very well. I does a correction which is much better than the one with a fieldmap.
>>
>>
>> I then try to run eddy on the combine data set. I merged the AP and PA acquisition in one 4D nifti and run the eddy command
>>
>> eddy --imain=4D_dti.nii.gz --mask=bmask_mask.nii.gz --index=index.txt --acqp=acqp.txt --bvecs=bvecs --bvals=bvals --out=test --session=session.txt
>>
>>
>> The output from eddy was a 4D volume almost identical the the input (it did not correct for the big AP PA distorsion). May be I need to call it with the --topup.
>>
>> But it is not clear to me on which data I should run topup before eddy : only on the B0 or on the all data set (including the direction)
> yes you need to specify the --topup switch for eddy, giving it the name of the output of your topup run.
>
> Jesper
>
ok but then I get the error

Terminate called after throwing an instance of ‘NEWMAT::IndexException’

this is the same as Stefano Marenco (23/10/2012) so I have to wait for 
the update... (?)


I test both (with the same error) :
1_ compute topup only on B0 or
2_ compute topup on the all

I just see in an other post that you advise the first one. but then how 
do you handel the index and acqp.txt parameters : in the eddy comand you 
specify the acqp and index for the full 4D data set (blip_up and 
blip_down concatenate) whereas topup was run only on the B0. (or in my 
case on several B0 since I have one B0 each 10 directions).
how do you specify the relation between the different B0 input of topup 
and the all data set input of eddy ?


thanks for your help

Romain