> Quoting KL Chen <[log in to unmask]>:
>
> > Dear experts,
> > we just start to use a new Siemens avanto machine to study DTI.
> However,> we can not find out the 6 diffusion gradient orientaiton
> sequence. Did
> > anyone know how to check the diffusion gardient orientation
> sequence from
> > dicom image ? or how can we know the sequence ?
>
> At least for those SIEMENS scanners I know, the relevant
> information is stored
> in the SIEMENS private part of the DICOM header. spm_dicom_headers
> reads this
> into hdr{j}.CSAImageInfo, from where you can retrieve it using the
> get_numaris4_val subfunction in spm_dicom_convert. Be aware that
> gradientdirection is given in scanner coordinates. If you want to
> use my dti toolbox
> for analysis, you have to transform this information into analyze
> coordinates,which means flipping the Y component of the direction
> vector.
> Hope this helps,
>
> Volkmar
>
hello all,
a little addition although maybe everybody knows this:
if you do not use slice orientations along one of the axes in scanner coordinates (i.e. if you use oblique slices) then you may also need to rotate your diffusion vectors along with your oblique slices - i am not sure maybe volkmar's toolbox already takes care of it. for example: suppose you acquire your slices rotated to the left. then a diffusion vector pointing along the z direction in the scanner coordinates (0,0,1) will not point perpendicular to the slices.
this matters if you want to display color-coded images or for fiber tracking. although the diffusion directions are still not rotated to the anatomy. so when you say left-right you can only refer to the image and not the person's head.
zoltan
p.s. as far as i know, GE rotates its diffusion gradients with the acquired slices, philips sometimes rotates them and sometimes not and siemens does not rotate them.
|