Output -> output(basil_name, output_name, subdir)
Outputs data from Basil, renaming it and regridding it to structural/standard space
OutputMasked -> output_masked(output_name, subdir, mask)
As above, but with masking. Used in PV correction to output parameters by tissue type
Calibrate -> calibrate(basil_name, output_name, m0val, multiplier, subdir)
Calibrate Basil parameter and output in all spaces
Report -> report(output_name, subdir, masktype)
Report mean value of parameter within specified named mask - used to report PV values
Normalise -> normalise(output_name, subdir, masktype)
Output parameter value normalised by dividing by mean within a mask- used in PVC
Normalise_var -> normalise_var(output_name, subdir, masktype)
Output parameter variance normalised by dividing by mean squared within a mask - used in PVC
Registration -> registration(regbase, transopt, distout)
Register an image to the structural image
Calibration -> calibration()
Calculate M0 value for calibration in reference region method
Dobasil -> do_basil()
Run BASIL
Dooutput -> do_output(subdir)
Do all the output to a specified output space subdir
Log
Warn
Info or warnings.
- Define usage text and functions as above -> Line 609/2104
- Parse options, display usage if required -> Line 873/2104
- Create directories, decide on output spaces, print welcome, set option defaults including WP mode -> Line 994/2104
- Preprocessing: -> Line 1275/2104
- Handle FSL_ANAT or structural image
- Remove vol 1 of [calib, cref, cblip)] image, MC and take mean
- MC asl data using calib as reference
- Tag control subtraction and reordering using ASL_FILE
- Get perfusion weighted image (mean over TIs)
- Registration of ASL data to structural image -> Line 1320/2104
- Segmentation of structural image and pull out sensitivity, PV maps -> Line 1385/2104
- Generate mask -> Line 1429/2104
- Distortion correction and repeat TC subtraction after correction -> *Line 1557/2104
- Set up BASIL options -> Line 1795/2104
- Run BASIL -> Line 1827/2104
- Re-run registration on perfusion map -> Line 1840/2104
- Do partial volume estimates -> Line 1881/2104
- Calibration -> Line 1968/2104
- Do output
- Re-run BASIL for PVC -> Line 2003/2104
- Do epoch analysis -> Line 2048/2104
- Final misc outputs and cleanup -> Line 2104/2104
The main oxasl program should end up as a series of calls to separate modules, each
of which has it's own command line tool as well as the direct API
|image | Basic image class for ASL data | DONE
|calib | Calibration (voxelwise and ref region) | DONE?
|mask | Mask generation | DONE?
|reg | Registration to structural data | DONE?
|preproc | Preprocessing (differencing, smoothing, moco?) | DONE?
|basil | Model fitting | DONE?
|struc | Handling structural data | DONE?
|discorr | Distortion correction | TODO
|pvc | Partial volume corrections | TODO
|epoch | Epoch analysis | TODO
|report | Human-readable report document | PARTIAL
Steps 1-6 are at least partially implemented and functional Step 7 (distcorr) is not yet implemented Steps 8-9 (basil) are implemented Steps 10-11 (re-registration and PVC) are not implemented Step 12 (calib) is implemented Step 13 (output) is incomplete Step 14 (re-fit for PVC) is not implemented Step 15 (epoch) is not implemented Step 16 is not implemented
How to pass data between modules? In particular data generated once should be re-used elsewhere, e.g.
- Brain-extracted versions of images
- Segmentations of structural image
Handle this using the Workspace class. Modules which need processed versions of data (e.g.
brain extracted structural) call the relevant module function to do this (e.g. preproc_struc).
Modules are responsible for:
- Not re-running themselves unnecessarily
- Not overwriting data provided explicitly (e.g. user-supplied brain extracted structural image)