Brain Imaging & Analysis Center

User Tools

Site Tools

You are here: biac » biac » analysis » resting_pipeline
biac:analysis:resting_pipeline

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision 		Previous revision
biac:analysis:resting_pipeline [2012/05/31 16:16]
admin [Step 6] 		biac:analysis:resting_pipeline [2023/02/23 18:43] (current)
Line 8: Line 8:
 Usage:  Usage: 
 resting_pipeline.py --func /path/to/run4.bxh --steps all --outpath /here/ -p func resting_pipeline.py --func /path/to/run4.bxh --steps all --outpath /here/ -p func
- 
  
 Program to run through Nan-kuei Chen's resting state analysis pipeline: Program to run through Nan-kuei Chen's resting state analysis pipeline:
Line 19: Line 18:
     5 - regress out WM/CSF     5 - regress out WM/CSF
     6 - bandpass filter     6 - bandpass filter
-    7 - produce correlation matrix from label file+    7 - do parcellation and produce correlation matrix from label file 
 +      * or split it up: 
 +         7a - do parcellation from label file 
 +         7b - produce correlation matrix [--func option is ignored if step 7b 
 +              is run by itself unless --dvarsthreshold is specified, and 
 +              --corrts overrides default location for input parcellation 
 +              results (outputpath/corrlabel_ts.txt)] 
 +    8 - functional connectivity density mapping 
  
  
Line 28: Line 35:
   --throwaway=4         number of timepoints to dis-regard from beginning of   --throwaway=4         number of timepoints to dis-regard from beginning of
                         run                         run
 +  --t1=/path/to/BXH     bxh ( or nifti ) file for the anatomical T1
   -p func, --prefix=func   -p func, --prefix=func
                         prefix for all resulting images, defaults to name of                         prefix for all resulting images, defaults to name of
Line 39: Line 47:
                         (1,3,5,2,4,6), up=ascending, down=descending,                         (1,3,5,2,4,6), up=ascending, down=descending,
                         even=interleaved (2,4,6,1,3,5) ).  Default is to read                         even=interleaved (2,4,6,1,3,5) ).  Default is to read
-                        this from input image, if available.+                        this from input BXH, if available.
   --tr=MSEC             TR of functional data in MSEC   --tr=MSEC             TR of functional data in MSEC
   --ref=FILE            pointer to FLIRT reference image if not using standard   --ref=FILE            pointer to FLIRT reference image if not using standard
Line 48: Line 56:
                         standard brain                         standard brain
   --refcsf=FILE         pointer to CSF mask of reference image if not using   --refcsf=FILE         pointer to CSF mask of reference image if not using
 +                        standard brain
 +  --refgm=FILE          pointer to GM mask of reference image if not using
 +                        standard brain
 +  --refbrainmask=FILE   pointer to brain mask of reference image if not using
                         standard brain                         standard brain
   --refacpoint=45,63,36   --refacpoint=45,63,36
                         AC point of reference image if not using standard MNI                         AC point of reference image if not using standard MNI
-                        brain                     +                        brain 
-  --betfval=0.5         f value to use while skull stripping. default is 0.5+  --betfval=0.4         f value to use while skull stripping. default is 0.4 
 +  --anatbetfval=0.5     f value to use while skull stripping ANAT. default is 
 +                        0.5
   --lpfreq=0.08         frequency cutoff for lowpass filtering in HZ.  default   --lpfreq=0.08         frequency cutoff for lowpass filtering in HZ.  default
-                        is .08hz+                        is .08hz.  highpass is fixed at .001hz.
   --corrlabel=FILE      pointer to 3D label containing ROIs for the   --corrlabel=FILE      pointer to 3D label containing ROIs for the
                         correlation search. default is the 116 region AAL                         correlation search. default is the 116 region AAL
Line 61: Line 75:
                         for the correlation search. default is the 116 region                         for the correlation search. default is the 116 region
                         AAL label txt file                         AAL label txt file
 +  --corrts=FILE         If using step 7b by itself, this is the path to 
 +                        parcellation output (default is to use 
 +                        OUTPATH/corrlabel_ts.txt), which will be used as input 
 +                        to the correlation. 
 +  --dvarsthreshold=THRESH 
 +                        If specified, this reprsents a DVARS threshold either 
 +                        in BOLD units, or if ending in a '%' character, as a 
 +                        percentage of mean global signal intensity (over the 
 +                        brain mask).  Any volume contributing to a DVARS value 
 +                        greater than this threshold will be excluded 
 +                        ("scrubbed") from the (final) correlation step.  DVARS 
 +                        calculation is performed on the results of the last 
 +                        pre-processing step, and is calculated as described by 
 +                        Power, J.D., et al., "Spurious but systematic 
 +                        correlations in functional connectivity MRI networks 
 +                        arise from subject motion", NeuroImage(2011).  Note: 
 +                        data is only excluded during the final correlation, 
 +                        and so will never affect any operations that require 
 +                        the full signal, like regression, etc. 
 +  --dvarsnumneighbors=NUMNEIGHBORS 
 +                        If --dvarsthreshold is specified, then 
 +                        --dvarsnumnumneighbors specifies how many neighboring 
 +                        volumes, before and after the initially excluded 
 +                        volumes, should also be excluded.  Default is 0. 
 +  --fdthreshold=THRESH  If specified, this reprsents a FD threshold in mm. 
 +                        Any volume contributing to a FD value greater than 
 +                        this threshold will be excluded ("scrubbed") from the 
 +                        (final) correlation step.  FD calculation is performed 
 +                        on the results of the last pre-processing step, and is 
 +                        calculated as described by Power, J.D., et al., 
 +                        "Spurious but systematic correlations in functional 
 +                        connectivity MRI networks arise from subject motion", 
 +                        NeuroImage(2011).  Note: data is only excluded during 
 +                        the final correlation, and so will never affect any 
 +                        operations that require the full signal, like 
 +                        regression, etc. 
 +  --fdnumneighbors=NUMNEIGHBORS 
 +                        If --fdthreshold is specified, then 
 +                        --fdnumnumneighbors specifies how many neighboring 
 +                        volumes, before and after the initially excluded 
 +                        volumes, should also be excluded.  Default is 0. 
 +  --motionthreshold=THRESH 
 +                        If specified, any volume whose motion parameters 
 +                        indicate a movement greater than this threshold (in 
 +                        mm) will be excluded ("scrubbed") from the (final) 
 +                        correlation step.  Volume-to-volume movement is 
 +                        calculated per pair of neighboring volumes from the 
 +                        three rotational and three translational parameters 
 +                        generated by mcflirt.  Motion for a pair of 
 +                        neighboring volumes is calculated as the maximum 
 +                        displacement (due to the combined rotation and 
 +                        translation) of any voxel on the 50mm-radius sphere 
 +                        surrounding the center of rotation.  Note: data is 
 +                        only excluded during the final correlation, and so 
 +                        will never affect any operations that require the full 
 +                        signal, like regression, etc. 
 +  --motionnumneighbors=NUMNEIGHBORS 
 +                        If --motionthreshold is specified, then 
 +                        --motionnumnumneighbors specifies how many neighboring 
 +                        volumes, before and after the initially excluded 
 +                        volumes, should also be excluded.  Default is 1. 
 +  --motionpar=FILE.par  If --motionthreshold is specified, then --motionpar 
 +                        specifies the .par file from which the motion 
 +                        parameters are extracted.  If you allow this script to 
 +                        perform motion correction, then this option is 
 +                        ignored. 
 +  --scrubop=SCRUBOP     If --motionthreshold, --dvarsthreshold, or 
 +                        --fdthreshold are specified, then --scrubop specifies 
 +                        the aggregation operator used to determine the final 
 +                        list of excluded volumes.  Default is 'or', which 
 +                        means a volume will be excluded if *any* of its 
 +                        thresholds are exceeded, whereas 'and' means all the 
 +                        thresholds must be exceeded to be excluded. 
 +  --powerscrub          Equivalent to specifying --fdthreshold=0.5 
 +                        --fdnumneighbors=0 --dvarsthreshold=0.5% 
 +                        --dvarsnumneigbhors=0 --scrubop='and', to mimic the 
 +                        method used in the Power et al. article.  Any 
 +                        conflicting options specified before or after this 
 +                        will override these. 
 +  --scrubkeepminvols=NUMVOLS 
 +                        If --motionthreshold, --dvarsthreshold, or 
 +                        --fdthreshold are specified, then --scrubminvols 
 +                        specifies the minimum number of volumes that should 
 +                        pass the threshold before doing any correlation.  If 
 +                        the minimum is not met, then the script exits with an 
 +                        error.  Default is to have no minimum. 
 +  --fcdmthresh=THRESH   R-value threshold to be used in functional 
 +                        connectivity density mapping ( step8 ). Default is set 
 +                        to 0.6. Algorithm from Tomasi et al, PNAS(2010), vol. 
 +                        107, no. 21. Calculates the fcdm of functional data 
 +                        from last completed step, inside a dilated gray matter 
 +                        mask
   --cleanup             delete files from intermediate steps?   --cleanup             delete files from intermediate steps?
 </code> </code>
Line 76: Line 181:
   * run fsl's slice time correction ( slicetimer )   * run fsl's slice time correction ( slicetimer )
   * if starting with a BXH header, you'll likely have the sliceorder field which will be used to create a custom sliceorder file to be used by fsl   * if starting with a BXH header, you'll likely have the sliceorder field which will be used to create a custom sliceorder file to be used by fsl
 +  * the default is to run bxh_slicetiming to extract a timing file from the BXH header
   * if this isn't present, then you'll need to define **--sliceorder** so that we can generate the file for you   * if this isn't present, then you'll need to define **--sliceorder** so that we can generate the file for you
     * "odd" is interleaved with odd, then even slice ordering: 1,3,5,2,4,6,etc     * "odd" is interleaved with odd, then even slice ordering: 1,3,5,2,4,6,etc
Line 81: Line 187:
     * "up" is ascending data, from the bottom up: 1,2,3,4,5,6     * "up" is ascending data, from the bottom up: 1,2,3,4,5,6
     * "down" is descending data, from the top down: 6,5,4,3,2,1     * "down" is descending data, from the top down: 6,5,4,3,2,1
 +  * this gets much more complicated with multi-band data, so using the default extraction is recommended 
  
 ==== Step 2 ==== ==== Step 2 ====
Line 89: Line 196:
   * the functional run is meaned across time with fslmaths, then bet is applied.  the resulting mask is then applied to the entire run of data   * the functional run is meaned across time with fslmaths, then bet is applied.  the resulting mask is then applied to the entire run of data
   * if you find that bet is doing a poor job finding the edges you can adjust the intensity threshold with **--betfval**   * if you find that bet is doing a poor job finding the edges you can adjust the intensity threshold with **--betfval**
-    * this is equivalent of **-f** when actually running bet.  the default is 0.5, but smaller values will be more conservative in finding the edges of the brain ( ie: larger mask ) +    * this is equivalent of **-f** when actually running bet.  the default is 0.4 (same as feat), but smaller values will be more conservative in finding the edges of the brain ( ie: larger mask ) 
 +  * if provided T1 anatomical is skull stripped. **--anatbetfval** is used to control intensity threshold, the default is 0.5 (same as feat)
 ==== Step 4 ==== ==== Step 4 ====
   * normalize the data using flirt   * normalize the data using flirt
Line 97: Line 204:
   * if your subject has already been normalized during standard pre-processing of other runs, please provide the flirt matrix from pre-stats with **--flirtmat** ( most likely example_func2standard.mat)   * if your subject has already been normalized during standard pre-processing of other runs, please provide the flirt matrix from pre-stats with **--flirtmat** ( most likely example_func2standard.mat)
     * this will apply the previously determined flirt matrix to your functional data instead of trying to calculate a new matrix based on the functionals.  the matrix from pre-stats was likely calculated using a high-resolution anatomical, also this will assure that your resting state runs are in the same space as the other runs from your subject.     * this will apply the previously determined flirt matrix to your functional data instead of trying to calculate a new matrix based on the functionals.  the matrix from pre-stats was likely calculated using a high-resolution anatomical, also this will assure that your resting state runs are in the same space as the other runs from your subject.
 +  * if you've provided at T1 anatomical image then this sequence is followed:
 +    * func-2-t1
 +    * t1-2-standard
 +    * flirt matrices are concatenated to create func-2-standard
 +  * if no T1 is provided, then the functional is used for the flirt normalization
  
 ==== Step 5 ====  ==== Step 5 ==== 
Line 106: Line 218:
 ==== Step 6 ====  ==== Step 6 ==== 
   * this step will band-pass filter data to remove high-frequency noise using custom python code   * this step will band-pass filter data to remove high-frequency noise using custom python code
-  * the default is 0.08 HZ+  * the default lowpass is 0.08 HZ 
 +  * highpass is fixed at .001 HZ
   * if you'd like to chose a different frequency, please use ** --lpfreq **   * if you'd like to chose a different frequency, please use ** --lpfreq **
 ==== Step 7 ====  ==== Step 7 ==== 
   * if defaults are used, then the aal_MNI_V4 label file is used to extract the average timeseries for each of the 116 regions   * if defaults are used, then the aal_MNI_V4 label file is used to extract the average timeseries for each of the 116 regions
 +  * if --dvarsthreshold or --motionthreshold are specified, then the ROI-averaged timeseries are further filtered ("scrubbed") by removing timepoints that exceed the given thresholds
   * cross correlation coefficients are found for the entire matrix of 116x116 regions   * cross correlation coefficients are found for the entire matrix of 116x116 regions
   * this step produces 4 files:   * this step produces 4 files:
Line 132: Line 246:
 * that is the zr_values from the subject.graphml displayed in matplotlib.  the cursor shows the intersecting region at the bottom, and if you click, time courses are displayed.  we're working on a release (cp - 3/30/12) * that is the zr_values from the subject.graphml displayed in matplotlib.  the cursor shows the intersecting region at the bottom, and if you click, time courses are displayed.  we're working on a release (cp - 3/30/12)
  
-For a beta version of the viewer:+For the rs-pipeline viewer:
 <code> <code>
-/usr/local/packages/biacpython/bin/beta_viewer.py --help+/usr/local/packages/biacpython/bin/rspipe_viewer.py --help
 Usage:  Usage: 
-beta_viewer.py --graphml /path/to/subject.graphml --stat zrvalue+rspipe_viewer.py --graphml /path/to/subject.graphml --stat zrvalue
  
 Program to display graphml output from resting_pipeline Program to display graphml output from resting_pipeline
Line 153: Line 267:
 </code> </code>
  
 +
 +==== Step 8 ==== 
 +  * Functional connectivity density mapping
 +  * Takes functional data from last step and calculates how connected they are to the voxels around them
 +  * uses ( --fcdmthresh and --refgm ) as the pearson r-value and gray matter mask
 +  * if defaults are used, then a dilated gray matter mask is used from FAST segmentation of MNI brain and a pearson r value of 0.6
 +  * Iteratively goes to all neighboring voxels and counts the number that have correlated signal until they are under the r threshold
 +  * adapted from Dardo Tomasi, PNAS(2010), vol. 107, no. 21. 9885–9890
 +  * resulting file with be "fcdm.nii.gz"; higher voxel values indicate more correlation from neighbors
 +
 +{{:biac:analysis:fcdm.png?direct&400|}}
 ===== Things to consider ===== ===== Things to consider =====
   * this was designed to be modular, so that you only need to run the steps you need   * this was designed to be modular, so that you only need to run the steps you need
Line 292: Line 417:
  
  
-Still under-development 
 3D VTK: 3D VTK:
  
Line 299: Line 423:
  
  
 +
 +
 +----
 +
 +===== Download Source =====
 +{{:biac:analysis:rsfmri_python.tgz|}}
 +  - source files assume you have a working install of FSL and all imported python modules
 +  - need a working install of the [[http://www.nitrc.org/projects/bxh_xcede_tools/|BIRN BXH/Xcede tools]]
 +  - will need to edit any paths that may be different at your install location ( FSL FAST segmentations of the MNI brain and base sets of ROIs )
 +  - **Chou et al. AJNAR(2012), May; 33(5): 833–838 **
 +  - fcdm algorithm adapted from **Dardo Tomasi, PNAS(2010), vol. 107, no. 21. 9885–9890**
  
  
  
  
biac/analysis/resting_pipeline.1338480972.txt.gz · Last modified: 2014/08/04 16:03 (external edit)

Page Tools

Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Share Alike 4.0 International
CC Attribution-Share Alike 4.0 International Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki