forked from Washington-University/HCPpipelines
-
Notifications
You must be signed in to change notification settings - Fork 0
Outputs to PreFreeSurfer
bstavel edited this page Mar 12, 2018
·
2 revisions
The purpose
- To average image repeats # PanLab does not use this step, since we only acquire one set of structurals
- To create a native, undistored strucutral volume space for the subject. Which means
- Distortion corrected for the gradient (scanner-specific, vednor-designed distoriton) and B0 distortions (fieldmap distortion) # We do not do Gradient Distortion Corection
- Rigidly realign to the axes of MNI space (but does not warp to MNI)
- Robust brain extraction
- Aligns T1 and T2
- Bias field correction
- Calculated from the square root of the product of the T1 and T2
- Warp to MNI
- Uses fnirt
Makes the following folders and files:
- T1w
* ACPCAlignment
- robustroi.nii.gz # is the T1 without the neck and lower head
- T1w_acpc.nii.gz # Resamples (if using a higher res template) and uses applywarp to linearly transform to MNI152 using output of acpc_final and a 6DOF approximation
- acpc_final.nii.gz # flirt used with cropped image 12DOF to linearly transform to MNI152 * BrainExtraction_FNIRTbased
- T1w_acpc_brain_mask # fslmaths to apply mni mask to T1_acpc (which is not nonlinearly-warped)
- T1w_acpc_to_MNI_roughlin.nii.gz # gets a first rough estimate of the linearly warp using 2mm template
- T1w_acpc_to_MNI_nonlin.nii.gz # warping rough linear and nonlinear image to the high res template * xfms Transformation matrices and warp fields * BiasFieldCorrection_sqrtT1wXT1w
- T1w_acpc_dc_brain.nii.gz # suffix of "_dc" indicates distortion corection. since we do not do distortion correction, T1w_acpc and T1w_acpc_dc are equivalent
- T1w_acpc_dc_restore.nii.gz # T1w_acpc image corrected by the bias_field image
- bias_raw.nii.gz # take the normalized sqrt image (but not smoothed image), mask out the csf and "extrapolate" to the entire FOV
- T1wmultT2w_brain_norm_s5.nii.gz # T1wmultT2w_brain_norm.nii.gz smoothed by a bias field smoothing sigma (in our case 5)
- T1wmultT2w_brain_norm.nii.gz # Literally sqrt(|T1w * T2w|)/mean(sqrt(|T1w * T2w|)) (our guess of bias field correction)
- T2w
* ACPCAlignment
- robustroi.nii.gz #
- T2w_acpc.nii.gz #
- acpc_final.nii.gz # * BrainExtraction_FNIRTbased
- T2w_acpc_brain_mask #
- T2w_acpc_to_MNI_roughlin.nii.gz #
- T2w_acpc_to_MNI_nonlin.nii.gz # * xfms Transformation matrices and warp fields *T2wToT1wDistortionCorrectAndReg
- Unknown must learn #
- MNINonLinear * xfms Transformation matrices and warp fields