Hi,
I'm not sure what the fieldmap images from a Bruker scanner are likely to contain, as the units and scaling are arbitrary and are likely to be different from Siemens (which is what the steps below are customised for).
In particular, are you sure that the phase is in radians? Note that I doubt it is in rad/s (radians per second) and if it were then you would skip steps 1 and 2 (and it would be unlikely to get two images). So I will assume that you meant radians rather than rad/s.
The -mul 1000 and -div 2 in step 2 are almost certainly wrong, as the output from step 1 should be in radians (you may need to scale the inputs to step 1 if the raw images are not in radians, which is easiest to check by looking at the range of the values and seeing if it is a 2*pi range). As you have a 1 ms echo time difference then you want to divide the phase difference by 0.001 seconds, which is the same as multiplying by 1000, but there is no need for the divide by 2.
The dwell time is absolutely key when you are applying the fieldmap with fugue (or epi_reg, which I'd recommend over fugue if you can get anatomical boundaries from a structural image and register to that - but in your case that might not be possible). This dwell time is the same as the effective echo spacing and you need to get this information from the functional image (EPI) protocol, but it is different from TR and TE. It is the time to readout a single line of k-space. If you know the total readout time (again, different from the TR) then it is related to the dwell time by: dwell_time = (total readout time)/N where N is the number of voxels in the phase encoding direction. If you also have in-plane acceleration (GRAPPA, SENSE, etc) then you also need to divide the dwell time by the acceleration factor.
Hopefully with the above modifications you can get things to work.
If not, let us know.
All the best,
Mark
> On 26 Oct 2017, at 02:28, SUBSCRIBE FSL S.J. Li <[log in to unmask]> wrote:
>
> Dear FSL experts
>
> I am using FSL prelude and FUGUE to correct my functional EPI image.
> Following is the data I have
>
> Dual gradient echo image with TR=50ms TE=2ms
>
> first echo time is 2ms
> 1.magnitude image 1-----mag_1
> 2.phase image 1 in rad/s-----pha_1
>
> second echo time is 3ms
> 3.magnitude image 2-----mag_2
> 4.phase image 2 in rad/s-----pha_2
>
> 5.functional image with TR=2000ms TE=20ms-----epi
>
> and the command line I used
>
> 1. unwrap the phase image
> prelude -a mag_1 -p pha_1 -o pha_1_unwrap
> prelude -a mag_2 -p pha_2 -o pha_2_unwrap
>
> 2. creat fieldmap
> fslmaths pha_2_unwrap -sub pha_1_unwrap -mul 1000 -div 2 fieldmap -odt float
>
> 3. correct the epi image
> fugue -i epi --dwell=0.001 --loadfmap=fieldmap -o result
>
> For the result I get didn't successfully correct the epi image, I am not sure the fieldmap and the dwell time I set is right in the command line.
> Or anyone have such experience on rat data before can give me a help to solve this question.
>
> Best regards!
>
> Louis
|