Output Files

Files Generated by the NPAIRS System

Output files gnerated by an NPAIRS analysis take on one of the following forms:

Filename Description

{id}.NPAIRS.{suffix} files that control an NPAIRS analysis

{id}.{model}.{iiii}.{suffix} analysis results from split data

{id}.{model}.ALL.{suffix} analysis results from all data

{id}.{model}.SUMM.{pattType}.{suffix} NPAIRS summary files

Filename	Description
{id}.NPAIRS.{suffix}	files that control an NPAIRS analysis
{id}.{model}.{iiii}.{suffix}	analysis results from split data
{id}.{model}.ALL.{suffix}	analysis results from all data
{id}.{model}.SUMM.{pattType}.{suffix}	NPAIRS summary files

where,

id = string identifying NPAIRS analysis

model = analysis method (e.g, GLM, CVA)

iiii = ith individual analysis (i = 1,...,2*Nsplits)

pattType = string identifying the pattern type (see table below)

suffix = file suffix (see tables below)

Common Pattern Types

Pattern Type Suffix Model Description

.tstat GLM T statistics volume

.cbeta GLM C*beta volume

.gis PCA eigenimages

.nwcgis CVA non-weighted (standard) canonical eigenimages

Pattern Type Suffix	Model	Description
.tstat	GLM	T statistics volume
.cbeta	GLM	C*beta volume
.gis	PCA	eigenimages
.nwcgis	CVA	non-weighted (standard) canonical eigenimages

NPAIRS Files

The NPAIRS files hold information on what data to analyze and how it is to be analyzed.

Suffix Description Type

.par NPAIRS parameter file ASCII

.grps (CVA) CVA group labels ASCII

.hdr VAPET header ASCII

.list volume list file ASCII

.vlfMap volume list file map ASCII

.vols data splitting ASCII

Suffix	Description	Type
.par	NPAIRS parameter file	ASCII
.grps (CVA)	CVA group labels	ASCII
.hdr	VAPET header	ASCII
.list	volume list file	ASCII
.vlfMap	volume list file map	ASCII
.vols	data splitting	ASCII

Summary Files

Summary files hold the final results from an NPAIR analysis. This includes such things as reproducibility, generalization, subject influence, and SPM's. The data in these files are collected across all the NPAIRS splits.

Suffix Description Type

.AVG average spatial pattern across all splits VAPET

.BW   (CVA) CVA between/within group ratios ASCII

.CC pattern correlation coefficients ASCII

.CV-TE   (CVA) canonical variates (test) VAPET

.CV-TR   (CVA) canonical variates (training) VAPET

.INFLU subject influence measurements ASCII

.POST   (CVA) posterior probablities ASCII

.ZS-AVG averages Z-score VAPET

.ZS-LOG Z-score log file ASCII

.ZS-S-HIST Z-score histograms (signal axis) VAPET

.ZS-N-HIST Z-score histograms (noise axis) VAPET

.ZS-S-BIN Z-score histogram bin values (signal axis) VAPET

.ZS-N-BIN Z-score histogram bin values (noise axis) VAPET

.HIST pattern histograms VAPET

.BIN pattern histogram bin values VAPET

.NOISE-STD standard deviation of noise axis ASCII

Suffix	Description	Type
.AVG	average spatial pattern across all splits	VAPET
.BW (CVA)	CVA between/within group ratios	ASCII
.CC	pattern correlation coefficients	ASCII
.CV-TE (CVA)	canonical variates (test)	VAPET
.CV-TR (CVA)	canonical variates (training)	VAPET
.INFLU	subject influence measurements	ASCII
.POST (CVA)	posterior probablities	ASCII
.ZS-AVG	averages Z-score	VAPET
.ZS-LOG	Z-score log file	ASCII
.ZS-S-HIST	Z-score histograms (signal axis)	VAPET
.ZS-N-HIST	Z-score histograms (noise axis)	VAPET
.ZS-S-BIN	Z-score histogram bin values (signal axis)	VAPET
.ZS-N-BIN	Z-score histogram bin values (noise axis)	VAPET
.HIST	pattern histograms	VAPET
.BIN	pattern histogram bin values	VAPET
.NOISE-STD	standard deviation of noise axis	ASCII

GLM Suffixes:
The GLM files are the files that get created by the GLM program. This includes the files that hold the design matrix, contrast vectors, and T-statistic volumes.

Suffix Description Type

.cbeta C*beta images VAPET

.contr contrast vector(s) ASCII

.defDmat design matrix definition vector ASCII

.dmat design matrix used in computations ASCII

.glmlog GLM log file ASCII

.gpart partitioning vector for design matrix ASCII

.icontr input contrast vector(s) ASCII

.idmat input design matrix ASCII

.info volume list file ASCII

.mean volume means ASCII

.pstd pooled STD from standard error image ASCII

.tstat T statistic images VAPET

Suffix	Description	Type
.cbeta	C*beta images	VAPET
.contr	contrast vector(s)	ASCII
.defDmat	design matrix definition vector	ASCII
.dmat	design matrix used in computations	ASCII
.glmlog	GLM log file	ASCII
.gpart	partitioning vector for design matrix	ASCII
.icontr	input contrast vector(s)	ASCII
.idmat	input design matrix	ASCII
.info	volume list file	ASCII
.mean	volume means	ASCII
.pstd	pooled STD from standard error image	ASCII
.tstat	T statistic images	VAPET

PCA Suffixes:
The PCA files are the files that get created by the PCA model. They include the files that hold the PC scores, PC eigenvalues and eigenvectors, etc.

Suffix Description Type

.eval eigenvalues ASCII

.gis eigenimages ASCII

.goff GIS offsets (SSM) ASCII

.gsf global scaling factors ASCII

.info volume list file ASCII

.mean volume means ASCII

.ssf principal component scores ASCII

.ssmlog PCA log file ASCII

.var total variance information ASCII

Suffix	Description	Type
.eval	eigenvalues	ASCII
.gis	eigenimages	ASCII
.goff	GIS offsets (SSM)	ASCII
.gsf	global scaling factors	ASCII
.info	volume list file	ASCII
.mean	volume means	ASCII
.ssf	principal component scores	ASCII
.ssmlog	PCA log file	ASCII
.var	total variance information	ASCII

CVA Suffixes:
The CVA files are the files that get created by the CVA model. They include the files that hold the canonical variates, canonical eigenimages, CVA eigenvalues and eigenvectors, etc.

Suffix Description Type

.can canonical variates (subset of '.cvs') ASCII

.cgis "weighted" canonical eigenimages VAPET

.chi chi-squared values from the CVA ASCII

.cvs canonical variates ASCII

.cvsAvg mean canonical variates ASCII

.eigval eigenvalues from the CVA ASCII

.eigvct eigenvectors from CVA ASCII

.group group labels ASCII

.info volume list file ASCII

.log CVA log file ASCII

.nwcgis non-weighted canonical eigenimages VAPET

.pcs principal components used in CVA ASCII

.r2 r^2 values between PC's and CV's ASCII

Suffix	Description	Type
.can	canonical variates (subset of '.cvs')	ASCII
.cgis	"weighted" canonical eigenimages	VAPET
.chi	chi-squared values from the CVA	ASCII
.cvs	canonical variates	ASCII
.cvsAvg	mean canonical variates	ASCII
.eigval	eigenvalues from the CVA	ASCII
.eigvct	eigenvectors from CVA	ASCII
.group	group labels	ASCII
.info	volume list file	ASCII
.log	CVA log file	ASCII
.nwcgis	non-weighted canonical eigenimages	VAPET
.pcs	principal components used in CVA	ASCII
.r2	r^2 values between PC's and CV's	ASCII

.par

The NPAIRS parameter file.

.grps

An ASCII file holding the group labels used in the CVA analyses.

IDL> ; ---- EXAMPLE ----
IDL> ; read in group labels
IDL> read_matrix, 'id.NPAIRS.grps', group, /FIX
IDL> ; group(i) = group label for the ith volume

.hdr

A VAPET header associated with the input data. This is used to get needed information such as voxel sizes and volume sizes.

IDL> ; ---- EXAMPLE ----
IDL> ; read in VAPET header
IDL> read_header, 'id.NPAIRS.hdr', hdr, /SILENT
IDL> cmvox = hdr.cmpix   ; voxel size in x,y,z
IDL> vsize = hdr.cmpix   ; volume size in x,y,z

.list

The volume list file. This is normally just a copy of the list file in the NPAIRS setup directory.

IDL> ; ---- EXAMPLE ----
IDL> ; read in the volume list file
IDL> ssm_rd_list, 'id.NPAIRS.list', info
IDL> session = info.session   ; scan session numbers
IDL> state = info.state       ; brain states

.vlfMap

An ASCII file holding the volume list file mappings. This file is copied directly from the '.vlfMap' setup file in the NPAIRS setup directory. The contents of this file are integer values - one for each volume - specifying the list file from which each volume came from. This file is needed when multiple list files are used in the creation of the NPAIRS setup files (see npairs_setup.pro). Since different list files hold volumes that may live in different directories, we need to know the list file each volume belongs to (see 'VolDir' and 'MskDir' keywords in the NPAIRS parameter file). If only one list file was used, then all the values in .vlfMap file are 1. If 2 list files were used, then the conents will contain 1's and 2's: a 1 meaning the volume came from the first list file, and a 2 meaning the volume came from the second list file.

IDL> ; ---- EXAMPLE ----
IDL> ; read in the volume list file map
IDL> read_matrix, 'id.NPAIRS.vlfMap', vlfMap

.vols

This is an ASCII file holding information on how to perform the disjoint data splits. The 2D array in the '.vols' file hold the volume numbers to be used in each NPAIRS split. The volume numbers are integer values that correspond to the values in the first column of the volume list file (see .list).

IDL> ; ---- EXAMPLE ----
IDL> ; read in split data
IDL> read_matrix, 'id.NPAIRS.vols', vid
IDL> vnums1 = vid(*,0)  ; volumes numbers for 1st split - 1st half
IDL> vnums1 = vid(*,1)  ; volumes numbers for 1st split - 2nd half

The 'vid' array is of size (#volumes in a split) X (2*#splits). 
vnums1 = vid(*,2*i+0)  ; volume numbers for ith split (1st half)
vnums2 = vid(*,2*1+1)  ; volume numbers for ith split (2nd half)

.can

This is an ASCII file holding the canonical variates for each CVA dimension.

IDL> ; ---- EXAMPLE ----
IDL> ; read in canonical variates from CVA of all data
IDL> read_matrix, 'id.CVA.ALL.can', cvs
IDL> cv1 = cvs(0,*)  ; canoncial variate scores for 1st CVA dimension

.cgis

A multiple volume VAPET file that holds the "weighted" canonical eigenimage volumes.

IDL> ; ---- EXAMPLE ----
IDL> ; read in canonical eigenimage volume for 1st CVA dimension
IDL> rd_volume, 'id.CVA.ALL.cgis 1', vol, hdr
IDL> survey_volw_w, V0=vol, CMVOX=hdr.cmpix

.chi

This is an ASCII file holding the Chi-squared results from the CVA.

IDL> ; ---- EXAMPLE ----
IDL> ; read in chi-squared info
IDL> read_matrix, 'id.CVA.ALL.chi', chisqr
IDL> chi = chisqr(0,*)  ; chi-squared statistic for each CVA dimension
IDL> chi = chisqr(1,*)  ; P-value associated with chi-squares
IDL> chi = chisqr(2,*)  ; chi-squared degrees of freedom

.cvs

This is an ASCII file holding the canonical variates for each CVA dimension. The conents of this file contain the data from the '.can' file plus canonical variates associated with any addition canonical eigenimage weights.

IDL> ; ---- EXAMPLE ----
IDL> ; read in canonical variates from CVA of all data
IDL> read_matrix, 'id.CVA.ALL.cvs', cvs
IDL> cv1 = cvs(0,2,*)  ; CV scores for 1st eigenimage weight and 3rd CVA dim

The size of the 'cvs' array is (#eigenimage wgts) X (#CVA dims) X (#vols).
cvs(i,j,k) = CV score for jth CVA dimension, kth volume, ith eigenimage weight.

.cvsAvg

This is an ASCII file holding the mean canonical variates for each CVA dimension and each canonical eigenimage weight.

IDL> ; ---- EXAMPLE ----
IDL> ; read in mean canonical variates from CVA of all data
IDL> read_matrix, 'id.CVA.ALL.cvsAvg', cvsAvg
IDL> cv1 = cvs(0,2,*)  ; mean CV scores for 1st eigenimage wgt and 3rd CVA dim

The size of the 'cvsAvg' array is (#eigenimage wgts) X (#CVA dims) X (#groups).
cvsAvg(i,j,k) = mean CV score for jth CVA dim, kth group, ith eigenimage wgt.

.eigval

This is an ASCII file holding eigenvalues from the CVA.

IDL> ; ---- EXAMPLE ----
IDL> ; read in CVA eigenvalues
IDL> read_matrix, 'id.CVA.ALL.eigval', eval

The size of the 'eval' vector is (#PC's use in CVA)

.eigvct

This is an ASCII file holding eigenvectors from the CVA.

IDL> ; ---- EXAMPLE ----
IDL> ; read in CVA eigenvectors
IDL> read_matrix, 'id.CVA.ALL.eigvct', evect
IDL> ev = evect(2,*)  ; eigenvector associated with 3rd CVA dimension

The size of the 'evect' array is (#CVA dimensions) X (#PC's used in CVA)

.group

An ASCII file holding the group labels used in the CVA analyses.

IDL> ; ---- EXAMPLE ----
IDL> ; read in group labels
IDL> read_matrix, 'id.CVA.ALL.group', group, /FIX
IDL> ; group(i) = group label for the ith volume

.info (CVA)

The volume list file.

IDL> ; ---- EXAMPLE ----
IDL> ; read in the volume list file
IDL> ssm_rd_list, 'id.CVA.ALL.info', info
IDL> session = info.session   ; scan session numbers
IDL> state = info.state       ; brain states

.log

The CVA log file.
<Top> <NPAIRS Files> <Summary Files> <GLM Files> <PCA Files> <CVA Files>

.nwcgis

A multiple volume VAPET file that holds the (standard) canonical eigenimage volumes.

IDL> ; ---- EXAMPLE ----
IDL> ; read in canonical eigenimage volume for 1st CVA dimension
IDL> rd_volume, 'id.CVA.ALL.nwcgis 1', vol, hdr
IDL> survey_volw_w, V0=vol, CMVOX=hdr.cmpix

.pcs

An ASCII file holding the PC numbers of the PC's used in the CVA analysis.

IDL> ; ---- EXAMPLE ----
IDL> ; read in PC numbers
IDL> read_matrix, 'id.CVA.ALL.pcs', pcNum
IDL> print, pcNum  ; print PC numbers

.r2

An ASCII file holding the correlation coefficients between the canonical variate scores and the principal components scores used to build them.

IDL> ; ---- EXAMPLE ----
IDL> ; read in r^2 values
IDL> read_matrix, 'id.CVA.ALL.r2', rSquared
IDL> r2 = rSquared(0,*)  ; correlations for 1st CVA dimension

The size of the 'rSquared' array is (#CVA dimensions) X (#PC's used in CVA).
rSquared(i,j) = correlation between ith CV and jth PC.

.eval

An ASCII file holding the eigenvalues from the PCA.

IDL> ; ---- EXAMPLE ----
IDL> ; read in PCA eigenvalues
IDL> read_matrix, 'id.PCA.ALL.eval', eval
IDL> print, eval   ; print the eigenvalues

The size of the 'eval' vector is #volumes.
eval(i) = eigenvalue corresponding to the ith PC.

.gis

A multiple volume VAPET file that holds the eigenimages from a PCA analysis.

IDL> ; ---- EXAMPLE ----
IDL> ; read in volume associated with 1st PCA dimension
IDL> rd_volume, 'id.PCA.ALL.gis 1', vol, hdr
IDL> ; vol(i,j,k) now holds the eigenimage value at location i,j,k
IDL> survey_vols_w, V0=vol, CMVOX=hdr.cmpix     ; display volume

.goff

An ASCII file holding the GIS offset values for an SSM analysis.

IDL> ; ---- EXAMPLE ----
IDL> ; read in GIS offsets
IDL> read_matrix, 'id.PCA.ALL.goff', gisOffsets
IDL> print, gisOffsets   ; print GIS offsets

The size of the 'gisOffsets' vector is equal to the number GIS's saved.
gisOffsets(i) = GIS offset for ith PC GIS

.gsf

An ASCII file holding the global scaling factors for an SSM analysis. The values in this file don't have meaning unless an SSM model was run.

IDL> ; ---- EXAMPLE ----
IDL> ; read in GSF's
IDL> read_matrix, 'id.PCA.ALL.gsf', gsf
IDL> print, gsf   ; print GSF

The size of the 'gsf' vector is #volumes.
gsf(i) = global scaling factor for ith volume

.info (PCA)

The volume list file.

IDL> ; ---- EXAMPLE ----
IDL> ; read in the volume list file
IDL> ssm_rd_list, 'id.PCA.ALL.info', info
IDL> session = info.session   ; scan session numbers
IDL> state = info.state       ; brain states

.mean (PCA)

An ASCII file holding various volume mean calculations performed on the transformed data matrix (see the 'Transf' keyword in the parameter file). The importance/usefulness of this file is minimal.

IDL> ; ---- EXAMPLE ----
IDL> ; read in the volume means
IDL> read_matrix, 'id.PCA.ALL.mean', mean
IDL> m1 = mean(*,0) ; pre-logged, pre-zeroed volume means
IDL> m2 = mean(*,1) ; post-logged, pre-zeroed volume means
IDL> m3 = mean(*,2) ; post-logged, post-zeroed volume means

.ssf

An ASCII file holding the principal component scores from the PCA. The '.ssf' suffix comes from SSM terminology, where the PC's are called Subject Scaling Factors.

IDL> ; ---- EXAMPLE ----
IDL> ; read in the PC scores
IDL> read_matrix, 'id.PCA.ALL.ssf', pcs
IDL> pc = pcs(*,i) ; PC scores for the ith principal component
IDL> pc = pcs(j,*) ; PC scores for the jth volume

.ssmlog

An ASCII file holding the PCA log file.
<Top> <NPAIRS Files> <Summary Files> <GLM Files> <PCA Files> <CVA Files>

.var

An ASCII file holding information on the total variance from the PCA. There are 3 numbers in this file: 1) number of volumes in analysis, 2) number of voxels, 3) total variance.

IDL> ; ---- EXAMPLE ----
IDL> ; read in total variance file
IDL> read_matrix, 'id.PCA.ALL.var', var
IDL> nVol = var(0) ; # volumes
IDL> nVox = var(1) ; # voxels
IDL> tVar = var(2) ; total variance of data

.AVG

A multiple volume VAPET file that holds the average spatial patterns (e.g CVA eigenimages) across all the NPAIRS analyses (2*#splits ). That is, for each split of the data 2 sets of analyses are run - one for each split half. These analyses generate spatial patterns (e.g, PCA eigenimages), and the .AVG summary file holds the average of these spatial patterns.

IDL> ; ---- EXAMPLE ----
IDL> ; read in volume associated with 1st CVA dimension
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.AVG 1', vol, hdr
IDL> ; vol(i,j,k) holds the average spatial pattern value at location i,j,k
IDL> survey_vols_w, V0=vol, CMVOX=hdr.cmpix     ; display volume

.BW

An obsolete file holding information about CVA between/within ratios and misclassification errors based on distances from CVA group means.

IDL> ; ---- EXAMPLE ----
IDL> ; read in BW statistics
IDL> read_matrix, 'id.CVA.SUMM.nwcgis.BW', bwStats

bwStats = fltarr(2*#Splits,#CVA dimensions,4)
bwStats(i,j,0) = between group
bwStats(i,j,1) = within group
bwStats(i,j,2) = between group / within group
bwStats(i,j,3) = % misclassfication error
for ith analysis (i = 0,...,2*#splits-1), jth CVA dimension

.CC

Holds the correlation coefficents (among other similarity measurements) between the spatial patterns generated by the disjoint NPAIRS splits. This is the file used for the reproducibility plots.

IDL> ; ---- EXAMPLE ----
IDL> ; read in the correlation coefficients for CVA canonical eigenimages
IDL> read_matrix, 'id.CVA.SUMM.nwcgis.CC', cc

The contents of this variable are:
cc( 0,i,j) = NPAIRS analysis number for 1st half of split
cc( 1,i,j) = NPAIRS analysis number for 2nd half of split
cc( 2,i,j) = number of voxels common to both patterns
cc( 3,i,j) = correlation coefficent  *** important one
cc( 4,i,j) = concordance correlation coefficient
cc( 5,i,j) = linear regression offset
cc( 6,i,j) = linear regression slope
cc( 7,i,j) = 99% relative signal distribution width (RSDW)
cc( 8,i,j) = 99% noise width
cc( 9,i,j) = 99% signal width
cc(10,i,j) = 95% relative signal distribution width (RSDW)
cc(11,i,j) = 95% noise width
cc(12,i,j) = 95% signal width

where,
i = ith NPAIRS splits (i = 0,...,#splits-1)
j = jth CVA dimension

.CV-TE

A VAPET volume that holds the test CV scores computed by projecting the test data onto the canonical eigenimages generated from the training data.

IDL> ; ---- EXAMPLE ----
IDL> ; read in test CV scores
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.CV-TE', cvs
IDL> cv = cvs(i,j,*)  ; CV scores for ith CVA analysis, jth CVA dimension
IDL> cv = cv(where(cv ne 0))  ; a 0 means the volume was not part of test set

The size of the CV array is (2*#splits) X (#CVA dimensions) X (#volumes).

.CV-TR

A VAPET volume that holds the training CV scores.

IDL> ; ---- EXAMPLE ----
IDL> ; read in CV scores
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.CV-TR', cvs
IDL> cv = cvs(i,j,*)  ; CV scores for ith CVA analysis, jth CVA dimension
IDL> cv = cv(where(cv ne 0))  ; a 0 means the volume was not part of test set

The size of the CV array is (2*#splits) X (#CVA dimensions) X (#volumes).

.INFLU

An ASCII file holding the subject (actually the scan session) influence measurements.

IDL> ; ---- EXAMPLE ----
IDL> ; read in influence measurements for c*beta GLM spatial patterns
IDL> read_matrix, 'id.GLM.SUMM.cbeta.INFLU', influ
IDL> ; get counts for 11th session, 2nd CVA dimension, using mean of 2 most
IDL> ; reproducible  patterns as reference, and higher correlation counts
IDL> count = influ(10,1,0,2))  

The size of the INFLU array is (#scan sessions) X (#dimensions) X 2 X 3.
The conents of the array variable are:
influ(i,d,j,0) = scanning session ID numbers
influ(i,d,j,1) = number of counts for lower correlation
influ(i,d,j,2) = number of counts for higher correlation
where, 
i = ith scanning session
d = dth dimension
j = 0 influence based on mean of 2 most reproducible patterns
j = 1 influence based on mean of all patterns

.POST

An ASCII file holding the posterior probabilities (among other stuff).

IDL> ; ---- EXAMPLE ----
IDL> ; read in posterior probabilities
IDL> read_matrix, 'id.CVA.SUMM.nwcgis.POST', post
IDL> ; get PP (using priors) for 5th analysis (3rd split, 1st half) and 21st volume
IDL> pp = post(4,20,0,0)

The size of the POST array is (2*#splits) X (#test vols) X 2 X (4+#CVA groups).
The conents of the array variable are:
post(i,j,p,0)   = posterior probability for true group
post(i,j,p,1)   = squared prediction error
post(i,j,p,2)   = predicted group based on maximum posterior probability
post(i,j,p,3)   = correct classification (0 = wrong, 1 = right)
post(i,j,p,4+g) = posterior probability of belonging to gth group
where,
i = ith CVA analysis (i=0,1 -> 1st split, i=2,3 -> 2nd split, etc)
j = jth test volume
p = 0, 1 (0 -> priors used, 1 -> priors not used)

.ZS-AVG

A multiple volume VAPET file that holds the average of the Z-score volumes across all the NPAIRS splits.

IDL> ; ---- EXAMPLE ----
IDL> ; read in mean Z-score volume associated with 1st CVA dimension
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-AVG 1', vol, hdr
IDL> ; vol(i,j,k) now holds the average Z-score value at location i,j,k
IDL> survey_vols_w, V0=vol, CMVOX=hdr.cmpix

.ZS-HI

A multiple volume VAPET file that holds the upper 5% values of the Z-score volumes across all the splits.

IDL> ; ---- EXAMPLE ----
IDL> ; read in "high" Z-score volume associated with 1st CVA dimension
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-HI 1', vol, hdr
IDL> ; vol(i,j,k) now holds the upper 5% Z-score value at location i,j,k
IDL> survey_vols_w, V0=vol, CMVOX=hdr.cmpix     ; display volume

.ZS-LO

A multiple volume VAPET file that holds the lower 5% values of the Z-score volumes across all the splits.

IDL> ; ---- EXAMPLE ----
IDL> ; read in "low" Z-score volume associated with 1st CVA dimension
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-LO 1', vol, hdr
IDL> ; vol(i,j,k) now holds the lower 5% Z-score value at location i,j,k
IDL> survey_vols_w, V0=vol, CMVOX=hdr.cmpix     ; display volume

.ZS-LOG

A log file that describes how the Z-score volumes were created.
<Top> <NPAIRS Files> <Summary Files> <GLM Files> <PCA Files> <CVA Files>

.ZS-N-HIST

A VAPET volume that holds the histograms of the Z-scores associated with the normalized noise axis patterns.

IDL> ; ---- EXAMPLE ----
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-N-HIST', hist
IDL> ; hist(i,j,k) = #hits for ith NPAIRS split, jth CVA dimension, kth bin
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-N-BIN', bin
IDL> ; bin(j,k) = bin location (noise value) for jth CVA dim, kth bin location
IDL> plot, bin(0,*), hist(4,0,*), PSYM=10 ; histogram for 5th split, 1st CVA dim

.ZS-S-HIST

A VAPET volume that holds the histograms of the Z-scores associated with the normalized signal axis patterns.

IDL> ; ---- EXAMPLE ----
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-S-HIST', hist
IDL> ; hist(i,j,k) = #hits for ith NPAIRS split, jth CVA dimension, kth bin
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-S-BIN', bin
IDL> ; bin(j,k) = bin location (signal value) for jth CVA dim, kth bin location
IDL> plot, bin(0,*), hist(4,0,*), PSYM=10  ; histogram for 5th split, 1st CVA dim

.ZS-N-BIN

A VAPET volume that holds the bin values associated with the noise axis Z-score histograms. This goes with ZS-N-HIST file.

IDL> ; ---- EXAMPLE ----
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-N-HIST', hist
IDL> ; hist(i,j,k) = #hits for ith NPAIRS split, jth CVA dimension, kth bin
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-N-BIN', bin
IDL> ; bin(j,k) = bin location (noise value) for jth CVA dim, kth bin location
IDL> plot, bin(0,*), hist(4,0,*), PSYM=10  ; histogram for 5th split, 1st CVA dim

.ZS-S-BIN

A VAPET volume that holds the bin values associated with the signal axis Z-score histograms. This goes with ZS-S-HIST file.

IDL> ; ---- EXAMPLE ----
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-S-HIST', hist
IDL> ; hist(i,j,k) = #hits for ith NPAIRS split, jth CVA dimension, kth bin
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.ZS-S-BIN', bin
IDL> ; bin(j,k) = bin location (signal value) for jth CVA dim, kth bin location
IDL> plot, bin(0,*), hist(4,0,*), PSYM=10  ; histogram for 5th split, 1st CVA dim

.HIST

A VAPET volume that holds the histograms for each of the generated spatial patterns (2*#splits of them).

IDL> ; ---- EXAMPLE ----
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.HIST', hist
IDL> ; hist(i,j,k) = #hits for ith analysis, jth CVA dimension, kth bin
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.BIN', bin
IDL> ; bin(j,k) = bin location (eigenimage score) for jth CVA dim, kth bin location
IDL> plot, bin(0,*), hist(4,0,*), PSYM=10  ; histogram for 5th split, 1st CVA dim

.BIN

A VAPET volume that holds the bin values (e.g, CVA eigenimage voxel values) associated with the histograms generated for each spatial pattern. This goes with HIST file.

IDL> ; ---- EXAMPLE ----
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.HIST', hist
IDL> ; hist(i,j,k) = #hits for ith NPAIRS split, jth CVA dim, kth bin
IDL> rd_volume, 'id.CVA.SUMM.nwcgis.BIN', bin
IDL> ; bin(j,k) = bin value for jth CVA dim, kth bin location
IDL> plot, bin(0,*), hist(4,0,*), PSYM=10  ; histogram for 5th split, 1st CVA dim

.NOISE-STD

An ASCII file holding the standard deviations of each of the noise axis patterns.

IDL> ; ---- EXAMPLE ----
IDL> ; read in the noise axis standard deviations
IDL> read_matrix, 'id.SUMM.CVA.nwcgis.NOISE-STD', std

std = fltarr(#splits,#dimensions)
std(i,j) = standard deviation of ith split, jth dimension

.cbeta

A multiple volume VAPET file that holds the C*beta volume(s). There is one volume for each contrast vector. The C*beta volumes are computed by performing the dot product of the contrast vector with each column in the GLM beta matrix.

IDL> ; ---- EXAMPLE ----
IDL> ; read in c*beta volume for the 1st contrast vector
IDL> rd_volume, 'id.GLM.ALL.cbeta 1', vol, hdr
IDL> ; vol(i,j,k) now holds the c*beta voxel value for location i,j,k
IDL> survey_vols_w, V0=vol, CMVOX=hdr.cmpix     ; display volume

.contr

An ASCII file holding the contrast vector(s) used to test the parameters in the beta matrix.

IDL> ; ---- EXAMPLE ----
IDL> ; read in contrast vector(s)
IDL> read_matrix, 'id.GLM.ALL.contr 1', contr
IDL> c1 = contr(*,0)  ; the 1st contrast vector

.defDmat

An ASCII file holding the design matrix definition vector. This is an 8 element integer vector that defines how the design matrix was created. The integer values correspond to columns in the volume list file. This file is somewhat obsolete.

IDL> ; ---- EXAMPLE ----
IDL> ; read in design matrix definition vector
IDL> read_matrix, 'id.GLM.ALL.defDmat', defDmat
IDL> d1 = defDmat(0)  ; studies
IDL> d2 = defDmat(1)  ; blocks
IDL> d3 = defDmat(2)  ; indicators
IDL> d4 = defDmat(3)  ; covariates
IDL> d5 = defDmat(4)  ; detrending
IDL> d6 = defDmat(5)  ; linear time
IDL> d7 = defDmat(6)  ; global
IDL> d8 = defDmat(7)  ; constant

.dmat

An ASCII file holding the design matrix used in the computation of the model pararameters (i.e, the beta matrix).

IDL> ; ---- EXAMPLE ----
IDL> ; read in design matrix
IDL> read_matrix, 'id.GLM.ALL.dmat', dMat
IDL> c1 = dMat(0,*)  ; 1st column in design matrix

.glmlog

The the GLM log file.
<Top> <NPAIRS Files> <Summary Files> <GLM Files> <PCA Files> <CVA Files>

.gpart

An ASCII file describing how the design matrix is partitioned. This file contains 7 numbers, with each number indicating the number indicator columns, number of covariate columns, number of linear time columns, number of detrending columns, number of block columns, number of global columns (e.g, 1 for ANCOVA), and the number of constant columns (either 0 or 1), respectively. This file is of little use.

IDL> ; ---- EXAMPLE ----
IDL> ; read in design matrix partition vector
IDL> read_matrix, 'id.GLM.ALL.gpart', gpart
IDL> n1 = gpart(0)  ; # indicator (condition) columns
IDL> n2 = gpart(1)  ; # covariate columns
IDL> n3 = gpart(2)  ; # linear time columns
IDL> n4 = gpart(3)  ; # detrending columns
IDL> n5 = gpart(4)  ; # block columns
IDL> n6 = gpart(5)  ; # global columns
IDL> n7 = gpart(6)  ; # constant columns

.icontr

An ASCII file holding the contrast vector(s). The contrast vectors in this file may differ from the contrast vectors in the '.contr' file. This difference is a result of a degenerate (non-full rank) design matrix, which the GLM program will fix by removing dependent columns in the matrix. As a result, the contrast vectors must be altered to account for the new, full rank design matrix. The altered contrast vectors, which are the ones actually used to compute the T statistics, are saved in the '.contr' file.

IDL> ; ---- EXAMPLE ----
IDL> ; read in contrast vector(s)
IDL> read_matrix, 'id.GLM.ALL.icontr', contr
IDL> c1 = contr(*,0)  ; the 1st contrast vector

.idmat

An ASCII file holding the design matrix. The design matrix in this file may differ from the design matrix in the '.dmat' file. This difference is a result of a degenerate (non-full rank) design matrix, which the GLM program will fix by removing dependent columns in the matrix. The new, full rank design matrix is then saved in the '.dmat' file, and is used in the computation of the beta matrix.

IDL> ; ---- EXAMPLE ----
IDL> ; read in design matrix
IDL> read_matrix, 'id.GLM.ALL.idmat', dMat
IDL> c1 = dMat(0,*)  ; 1st column in design matrix

.info (GLM)

The volume list file.

IDL> ; ---- EXAMPLE ----
IDL> ; read in the volume list file
IDL> ssm_rd_list, 'id.GLM.ALL.info', info
IDL> session = info.session   ; scan session numbers
IDL> state = info.state       ; brain states

.mean (GLM)

An ASCII file holding the volume means. The volume means are used to proportionally scale (VMN) each volume (row) in the data matrix, or as the values in the ANCOVA column in the design matrix.

IDL> ; ---- EXAMPLE ----
IDL> ; read in volume means
IDL> read_matrix, 'id.GLM.ALL.mean', mean
IDL> m = dMat(i)  ; mean of ith volume

.pstd

An ASCII file holding the pooled standard deviations computed from the standard errors. There is one pooled STD for each contrast vector. The pooled STD could be used to compute T volumes based on pooled STD's instead of voxel specific error measurements.

IDL> ; ---- EXAMPLE ----
IDL> ; read in pooled standard deviations
IDL> read_matrix, 'id.GLM.ALL.pstd', pstd
IDL> std = pstd(0)  ; pooled standard deviation for 1st contrast vector

.tstat

A multiple volume VAPET file that holds the T statistic volume(s). There is one volume for each contrast vector. The T statistic volumes are computed by dividing the C*beta volume by the standard error, where C is a contrast vector, and beta the 2D array of estimated model parameters.

IDL> ; ---- EXAMPLE ----
IDL> ; read in T volume for the 1st contrast vector
IDL> rd_volume, 'id.GLM.ALL.tstat 1', vol, hdr
IDL> ; vol(i,j,k) now holds the T value for location i,j,k
IDL> survey_vols_w, V0=vol, CMVOX=hdr.cmpix     ; display volume

id	= string identifying NPAIRS analysis
model	= analysis method (e.g, GLM, CVA)
iiii	= ith individual analysis (i = 1,...,2*Nsplits)
pattType	= string identifying the pattern type (see table below)
suffix	= file suffix (see tables below)