No need to rescan. You should simply get a mask from the diffusion data, not the T1.
Cheers
Stam
> On 7 May 2016, at 10:38, Vasileios Kokkinos <[log in to unmask]> wrote:
>
> Hi Stam,
>
> Thanx for the swift response.
>
> Already done that through BET and the error persists. But it seems that it needs an exact match, right? The whole-brain T1 mask should fully correspond to the diffusion acquisition brain. Even a single slice off can produce a mismatch. This wasn't the case in Neuro3D, as it would coregister structural and tensor data without 100% overlap. Obviously scanning has to adapt to FSL as well.
>
> I'll rescan with whole-brain DTI and give it a go anew.
>
> Thanx.
> Vasileios.
>
>
> On 07/05/2016 12:21 μμ, Stamatios Sotiropoulos wrote:
>> Hi Vasileios
>>
>> Yes you need a mask that matches your diffusion data. You can get that by running bet on the data. Have a look at the online documentation for more details.
>>
>> Cheers
>> Stam
>>
>>
>>
>>
>>> On 7 May 2016, at 09:50, Vasileios Kokkinos <[log in to unmask]> wrote:
>>>
>>> Hi everyone,
>>>
>>> I have been working for years with Siemens' Neuro3D for tractography at the clinic and only recently kicked off with FSL in order to process data from other scanners. So I stumbled upon this problem, and I'd appreciate any help, if anyone has come across it as well.
>>>
>>> I get a kill during the estimation process by BEDPOSTX (with default settings). At some point during the post-processing stage, the process halts, and pops up the error message: "Errors: child killed: kill signal" and prompts to the log file, where the error is:
>>>
>>> Image Exception : #3 :: Mask of different size used in matrix()
>>> An exception has been thrown
>>> Mask of different size used in matrix()
>>>
>>> Could it be the fact that the diffusion acquisition is not whole brain and crops a bit of cortex on the top, thereby making the structural mask (derived from a T1 whole-brain scan) incompatible?
>>>
>>> Could it be that the diffusion images are weighted (2)? Or could it be the effect of other acquisition parameters that FSL regards are pre-set?
>>>
>>> The acquisition settings for the DTI are below:
>>> TR:8000, TE: 94, 50 slices, 2x2x2, A->P phase encoding, dist fact 0%, 6 averages 1 concatenation in a single series, iPat GRAPPA acc.f: 2, axial orientation, MDDW diffusion mode, Diff weighting 2, b1:0, b2:1000.
>>>
>>> Any idea or help would be most appreciated.
>>>
>>> Vasileios Kokkinos
|