Good morning Mike,
My responses are in the text:
>My first question is, I think I should perform fieldmap correction and then do preprocessing (realignment, coregistration, normalization, ...etc). Is it correct??
In fact, you have to used the fieldmap toolbox to generate the VDM5 file (voxel displacement mask). Then you apply the correction in the same time as the realignment by using the function "Realignment and Unwarp". When you give a VDM file to this function, it will correct two things in your fMRI data:
- the Movement-by-susceptibility-distortion interaction
- voxel displacement due to B0 inhomogeneity (mainly in the orbitofrontal area)
>Then, in spm8, I open batch --> tools --> FieldMap. Below is my input data by order:
>In "Presubtracted phase and Magnitude data":
Yes, in the case of Siemens, the scanner usually compute automatically the subtraction between the two phase map. This should be the right option for you.
>Phase image: phase image (the result from subtraction of two images)
>Magnitude image: the magnitude image with the shorter echo time
Ok
>Fieldmap defaults: Defaults values
>Echo time: 10.00 12.46
Ok, just a remark for the acquisition. Keep in mind that to have good result with the fieldmap, you should choose two echo time where the water and fat are in phase. Ask your physicist that it is the case. If you don't know how to calculate these value, I explained it in a previous post: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1310&L=spm&P=R69493&1=spm&9=A&J=on&X=4EF6165943D0E5BD27&d=No+Match%3BMatch%3BMatches&z=4
>Mask brain: should I select "no brain masking" or, if I select "Mask brain", what should I select to enter??
In fact, these option don't make really a difference. I tested them and both gave similar results. Usually I use the option Mask brain but it's up to you.
>Blip direction: -1 (because the phase-encoding direction of my fieldmap is from right to left. Is it correct??)
The blip direction don't refer to the direction of phase-encoding of the fieldmap but of your EPI images. Blip of -1 means that the phase-encoding was in anterior to posterior direction. I think that (-1) is a good choice because I have the same with my Siemens PRISMA. If you want to be sure, you can test both solution (-1 and 1). Write the unwarped EPI and make a comparison between the two unwarped files and the anatomic scan with check Reg. Look the orbitofrontal voxels in which version their position is the more coherent with the anatomical scan.
>Total EPI readout time: It is weird that the terminal of Prisma scanner (i.e. in "sequence" page) does not show the value of echo spacing. How come?? The fieldmap is 128x128, so I think the total EPI readout time should be >128*(echo spacing time).
You have to calculate it on the base on the echo-spacing value, number of phase echo (which correspond to you matrix size) and phase acceleration factor (iPAT factor for Siemens). You need values from your EPI sequence (fMRI sequence) and not from the fieldmap sequence! These three values could be found in the terminal. The formula is:
EPI readout time = (number of phase echo * echo-spacing)/(phase acceleration factor)
>EPI-based fieldmap: non-EPI based
Ok, non-EPI based because you have acquired a fieldmap
>Jacobian modulation: do not use
Ok
>EPI sessions/select EPI to unwarp: I randomly select a single EPI image (called EPI_1) from my EPI data
No, you need to select the EPI file which is going to be used as reference in realignment. Select the first EPI that you will enter in realignment and use it as reference (default option in realignment and unwarp function)
>Match VDM to EPI?: match VDM (no now field map data would be coregistered to EPI_1, right?)
Yes, it coregister your VDM file with you reference EPI.
>Write unwarped EPI?: Write unwarped EPI
No, actually you don't need to write it because it should be write again when you are going to use the realignment and unwarp function.
>Select anatomical image for comparison: I select the T1 image of that subject. Am I correct??
Yes, but this step is facultative. It is just for visual control. I suggest you to put None here.
>Match anatomical image to EPI?: match anat
NO, you should not match to anat if it is a different position as the reference EPI (which will be you case because you have not yet coregister the T1 with your reference EPI). The important is that the VDM is coregister with your reference EPI !
>In “Apply VDM”:
>Images: should I select all the EPI images I want to unwarp using field map data??
>Fieldmap: the output vdm* file created from the fieldmap toolbox described above
>Reslice Options/Reslice which images?: should I select "all images + mean images" or "all images"? what's the difference between both? I think because in the following preprocessing procedures (1) I will use mean images as >the reference of coregistration for T1 and (2) all images should be corrected by field map for the following preprocessing, I should select "all images + mean images". Am I correct??
>Masking: should I select Mask or don't mask? what's the difference??
NO, I suggest you to use the function "realign and unwarp" to apply the VDM instead of the "apply VDM" of the batch editor.
Otherwise, if you use "Apply VDM" of batch editor, you would need to realign all your EPI before. This tool will apply only the VDM correction. The realign and unwarp offer you to step more: realignment and the Movement-by-susceptibility-distortion interaction correction.
Best regards,
Michael
|