An Automatic and Subject-specific Method for Locus Coeruleus Localization and BOLD Activity Extraction
H He, L Hong, P Sajda
Proceedings of the International Society for Magnetic Resonance in Medicine 2021
By Hengda He - LIINC - Columbia University
June 20th 2021
1. Subject TSE image space and MNI standard space registration
1) TSEandMNI_registration/FSLregistration_TSE_1mm.sh
Input: TSE_image, T1_image, standard_image
a. Rigid registration of TSE image to T1_head
b. T1_brain non-linear registration to MNI_standard_brain
Output: TSE2structhead.mat, T1brain2strandardbrain_warp.nii.gz, TSE2strandardbrain_warp.nii.gz
2) TSEandMNI_registration/LCmask_MNI2SUB.sh
Input: LC_atlas
Warp LC_atlas into subject structural space.
Output: LC_template in Structspace
2. Hybrid method to locate LC in functional space, and extract time series
1) LC_HybridMethod/main.sh
2) LC_HybridMethod/Subject_Classification_Criterion.m
Input: TSE_image_in_T1Structuralspace, LC_template in StructuralSpace
Move voxels in 1SD template to local max with constraints
Output: Statistics and classification of subjects into groups for Hybrid method.
3) LC_HybridMethod/FSLregistration_co_reg_Stru2EPI_applyLCTSE_1SD_BBR.sh
Input: TSE in StructuralSpace, LC_template in StructuralSpace
a. BBR registration of high_res_head and T1_image_head
b. Mask TSE image with LC_1SD_template in Structural Space
c. EPI_head rigid registration to high_res_head
d. Transform Masked_TSEintensities_inLC1SD_Template from structural space to functional space
Output: TSE_images masked with LC_template_1SD in functional space
4) LC_HybridMethod/FSLregistration_co_reg_Stru2EPI_applyLCTSE_2SD_BBR.sh
Output: TSE_images masked with LC_template_2SD in functional space
5) LC_HybridMethod/FSLregistration_co_reg_Stru2EPI_applyLCtemp.sh
Output: LC_1SD_template_binary_mask trilinear interpolated into functional space
6) LC_HybridMethod/LCmask_thresholdProb.m
Input: trilinear interpolated LC TSEintensities or mask in functional space
Output: LC location in functional space with 2 Vox version and 6 Vox version
7) LC_HybridMethod/fsl_preprocessing_feat.sh
Input: fMRI
No spatial smoothing. Slice timing correction, motion correction, temporal filtering
Output: preprocessed fMRI
8) LC_HybridMethod/FSLregistration_co_reg_Stru2EPI_apply4thV_FS.sh
Input: Freesurfer segmentation, T1, highres and EPI images
a. Extract 4th ventricle mask in Freesurfer space
b. BBR registration of Freesurfer space and high res space
c. Rigid registration of EPI functional space and high res space
d. Transform 4th ventricle binary mask from Freesurfer space to functional space
9) LC_HybridMethod/extract_timeseries.sh
Input: LC location and probability in functional space (2 Vox and 6 vox), preprocessed fMRI.
a. scale fMRI into percent signal change
b. Extract time series from 4th ventricle
c. GLM to regress out motion parameters (standard and extended) and 4th ventricle time series.
d. Extract time series in LC locations weighted with LC location probabilities from GLM residual.
Output: LC time series (2 vox and 6 vox versions)