History of AFNI updates  

|
Jim Eliassen
December 12, 2002 10:32AM
Hi Eric,

We typically cluster sub-briks in pairs, a coefficient and its associated t-statistic. The following script loops through a set of sub-briks stepping two sub-briks per loop. It uses the "brik_ct_plus" sub-brik, the odd one of the pair of sub-briks, to calculate the cluster and t-statistic cutoff, then applies this mask to the "brik_ct" sub-brik (the even numbered one of the pair that precedes the t-statistic). If you want to use an F-statistic sub-brik to threshold all other sub-briks you could move the cluster calculating step to a position outside and before the loop begins, then use that as a mask for all the sub-briks you want to mask. Just remember to step by the right increment for your data

-jim



#!/bin/csh -f

nice +20


@ brik_ct_max = 20 ## CHECK MAX NUMBER IN YOUR DATA##

#INITIALIZE THE COUNTER VARIABLES
@ brik_ct = 0 #PICK THE SUB-BRIK YOU WANT TO START WITH
@ brik_ct_plus = 0

while ($brik_ct < $brik_ct_max)

@ brik_ct_plus = $brik_ct + 1

#MASKS OUT NON-BRAIN
3dcalc -fscale \
-a "2factor+tlrc[${brik_ct_plus}]" \
-b ../coverage/12sub_coverage_mask+tlrc \
-expr "a*(b/b)" \
-prefix subbrick_${brik_ct_plus}_covmask

#USES OUTPUT OF PREVIOUS STEP
#T- OR F-VALUE CLIPPING AND CLUSTERING OF STATISTICAL SUB-BRIK
### CHECK FOR CORRECT T- OR F-VALUE###
3dmerge -1clip 2.901 \
-1clust 3 594 \
-prefix subbrick_${brik_ct_plus}_tclip \
subbrick_${brik_ct_plus}_covmask+tlrc

#T-/F-VALUE AND CLUSTER CLIPPED SUB-BRIK MASKS COEF. SUB-BRIK
3dcalc -fscale \
-a "2factor+tlrc[${brik_ct}]" \
-b subbrick_${brik_ct_plus}_tclip+tlrc \
-expr "a*(b/b)" \
-prefix subbrick_${brik_ct}_tclip

#STEP BY TWO TO BEGIN WITH NEXT T-STAT SUB-BRIK
@ brik_ct = $brik_ct + 2

end #while

#BUCKET FROM 0 to brik_ct_max - 1 IN COEF/T-STAT SUB-BRIK PAIRS
3dbucket -prefix clip_005_05_covtclip \
-fbuc subbrick_t${tp}_0_tclip+tlrc \
-fbuc subbrick_t${tp}_1_tclip+tlrc \
-fbuc subbrick_t${tp}_2_tclip+tlrc \
-fbuc subbrick_t${tp}_3_tclip+tlrc \
-fbuc subbrick_t${tp}_4_tclip+tlrc \
-fbuc subbrick_t${tp}_5_tclip+tlrc \
-fbuc subbrick_t${tp}_6_tclip+tlrc \
-fbuc subbrick_t${tp}_7_tclip+tlrc \
-fbuc subbrick_t${tp}_8_tclip+tlrc \
-fbuc subbrick_t${tp}_9_tclip+tlrc \
-fbuc subbrick_t${tp}_10_tclip+tlrc \
-fbuc subbrick_t${tp}_11_tclip+tlrc \
-fbuc subbrick_t${tp}_12_tclip+tlrc \
-fbuc subbrick_t${tp}_13_tclip+tlrc \
-fbuc subbrick_t${tp}_14_tclip+tlrc \
-fbuc subbrick_t${tp}_15_tclip+tlrc \
-fbuc subbrick_t${tp}_16_tclip+tlrc \
-fbuc subbrick_t${tp}_17_tclip+tlrc \
-fbuc subbrick_t${tp}_18_tclip+tlrc \
-fbuc subbrick_t${tp}_19_tclip+tlrc

rm subbrick_t${tp}_*
Subject Author Posted

clustering and merging bucket data sets

Eric December 11, 2002 03:47PM

Re: clustering and merging bucket data sets

Jim Eliassen December 12, 2002 10:32AM

Re: clustering and merging bucket data sets

Eric December 12, 2002 04:54PM

Re: clustering and merging bucket data sets

Jim Eliassen December 13, 2002 01:27PM