Table 4.
Automated image processing pipelines.
| Name | Description | Products | Core Methods | Run Time |
|---|---|---|---|---|
| Protocol validation | Inspects DICOM headers to verify image acquisition | Validation Report | XNAT | 1m |
| Defacing | Blurs areas of structural images containing facial features. | DICOM w/ modified voxels | Milchenko & Marcus | 1m/series |
| NIFTI conversion | Generates NIfTI-formatted versions of acquired images | NIFTI | dcmtonii | 5m/series |
| Phantom QC | Calculates RMS stability, drift, mean value, and SNR | QC Report | BIRN | 15m |
| fMRI QC | Calculates mean, variance, skewnewss and kurtosis of DVAR | QC Report | FSL | 15m |
| Structural MR QC | Detects blurring, edge coherence, and SNR | QC Report | python | 5m |
| Pre-surface generation | Removes spatial artifacts and distortions from structural scans, co-registers them to common atlas space. | Undistorted native and MNI volumes | FSL/python | 8h |
| Surface generation | Generates cortical surfaces and segmentations. | Native Surfaces, Segmentations | Freesurfer, FSL | 24h |
| Post-surface generation | Generates Workbench-ready data files and myelin maps. | Workbench spec files | Workbench,FSL | 4h |
| fMRI volumetric processing | Removes spatial distortion, motion correction, and co-registers structural scans to common atlas space. | MNI registered fMRI volumes | Freesurfer, FSL, python | 4h |
| fMRI surface processing | Maps fMRI volume timeseries to surfaces and creates standard grayordinates space. | CIFTI dense timeseries in grayordinates | Workbench, FSL | 4h |
| Diffusion* | Intensity normalization, EPI distortion removal, eddy-current distortion removal, motion/gradient-nonlinearity correction, structural registration | Diffusion Data in Structural Space | FSL, Freesurfer, python | 36h |
| Fiber Orientation Modeling* | Bayesian estimation of diffusion parameters obtained using sampling techniques | Probabilistic diffusion orientations | FSL | TBD |
| MEG/EEG* | TBD | TBD | Fieldtrip | TBD |
| Connectivity* | TBD | TBD | Workbench/FSL | TBD |
| FIX* | Run ICA and classify ICA components with removal of noise-based components | rfMRI Denoising | FSL/matlab/R | 14h |
| Task fMRI Subject Analysis* | Single subject task fMRI level 1 and 2 GLM analysis | Workbench/FSL | 20m |
The HCP's pipelines are implemented using XNAT's Pipeline Service. Run time is shown in minutes based on execution on a XXX system. Pipelines denoted with an asterisk are currently in development and run times are estimated