Brain Activation and Morphological Mapping: A Tutorial

1
Brain Activation and Morphological Mapping A
Tutorial

• Synopsis
• Overview processing order
• Time-series analysis f-bamm
• Group analysis g-bamm
• Second level analysis x-bamm

BAMM
2
Overview
http//www.psychiatry.cam.ac.uk/BMU/BAMM/PUBLICATI
ONS/store/UserGuide-fbamm.doc http//www.psychiat
ry.cam.ac.uk/BMU/BAMM/PUBLICATIONS/store/UserGuide
-xbamm.doc
BAMM
3
Overview
Task
haemodynamics
time-series
design matrix (model)
test statistic
Activation map
Activation map
Activation map

Difference map
Group activation map
BAMM
4
Overview
Getting started Edit .cshrc and add
lines setenv BAMM /home/js369/apps/BAMM-linux se
t path (path BAMM/bin) or setenv BAMM
/home/js369/apps/BAMM-solaris set path (path
BAMM/bin)
BAMM
5
Overview
Getting started Edit .bashrc and add
lines export BAMM /home/js369/apps/BAMM-linux e
xport PATHPATHBAMM/bin or export BAMM
/home/js369/apps/BAMM-solaris export
PATHPATHBAMM/bin
BAMM
6
f-bamm input
Directory structure for data organisation. UNC
format images
study/
images/
XBAMM/
GBAMM/
files
BAMM
7
f-bamm input
Type I design matrix Forsynchronous stimulation
acquisition. Each row is corresponds to an
image, each column to a contrast. Type II
design matrix Forasynchronous stimulus
acquisition or event related designs. Events
are denoted by onset and duration.
BAMM
8
f-bamm input
Type I design matrix format 1 0 0 0 1 0 0
0 .. 0 0 0 0 0 0 0 0 .. -2 1 0 1
Single space between columns. Any number can be
a contrast. No trailing spaces. unix2dos for
conversion
BAMM
9
f-bamm input
Type II design matrix format 1 0 0.4,3.5
2.5,2.5 17.5,1.0 0 1 20.0,1.1 34.0,0.6
Contrasts can be any value. First number is event
start time. Second number is duration. Use comma
and single space. No trailing spaces.
BAMM
10
f-bamm input

• .param files containing information to process
  data from a single case.
• /home/user/study/images/case1/bloggs_j.001
• /home/user/study/designfile
• 50
• 1
• WAVELET
• No trailing spaces.

BAMM
11
f-bamm input
Automated analysis A script is necessary to run
several cases !/bin/csh fbamm.csh lt
case1.param fbamm.csh lt case2.param fbamm.csh lt
case3.param
BAMM
12
f-bamm output
Important files (in case directories) .FST? UNC
images of the observed test statistic the
responses the paradigm produced. .RC?-? UNC
images of responses produced under conditions of
the null-hypothesis. bamm.log Contains all the
runtime info.
BAMM
13
f-bamm output
.ACT?.ppm Voxelwise test of significance (at
eppi) .CON?.ppm Clusterwise test of significance
(at ecpi)
Unsigned ..or signed?
BAMM
14
f-bamm clearing up
fbammclean Subject Prefix /home/user/study/image
s/case1/abcd.001 This process is destructive!!
Continue(y/n)? y Destroy motion correction
parameters(y/n)? n
BAMM
15
g-bamm input
Bring together all data from a group to assess
generic or median activation. In doing so, images
are registered into Talairach Standard
Space. .param contains the information to
generate group map /home/user/study/GBAMM/gbamm.
list /home/user/BAMM/Templates/Linux/Functional/Fu
ncTemplateA /home/user/study/GBAMM/output 50 1
BAMM
16
g-bamm input
.list contains all the subjects to be included
in the group map. /home/user/study/images/case1/c
ase_1.004 /home/user/study/images/case2/case_2.005
/home/user/study/images/case7/case_7.002 /home/us
er/study/images/subject77/001.bloggs_j
BAMM
17
g-bamm input
Choose the template that matches your
acquisition. For data from WBIC Bruker 3T use
Template3T. Choose either BAMM/Templates/Linux/
Functional/Template3T BAMM/Templates/Solaris/Func
tional/Template3T depending on architecture.
BAMM
18
g-bamm input
Automated analysis Update the script !/bin/bas
h fbamm.csh lt case1.param fbamm.csh lt
case2.param fbamm.csh lt case3.param gbamm.csh lt
gbamm.param
BAMM
19
g-bamm output
Important files (in output directory) .GST These
are UNC images of the observed test statistic
resliced and registered to template. Use these
for second-level (x-bamm) analysis. Groupmap.gmsk
UNC image showing registration
bamm.log Contains all the runtime info
BAMM
20
g-bamm output
.GVX?tal.ppm Voxelwise significance test
(eppi) .GCA?tal.ppm Clusterwise significance test
(ecpi)
BAMM
21
g-bamm output
.GGA?tal.info Text file detailing the positions
of clusters by identifying the geometric centre
and volume. Signed responses are separately
considered. 3D clusters are described difficult
to interpret. 2D clusters are described on each
slice independently. Regions of activation can
then be tracked from one slice to anther.
BAMM
22
g-bamm clearing up
gbammclean List of subjects gbamm.list This
process is destructive!! Continue(y/n)? y Destroy
mapping parameters(y/n)? n
BAMM
23
x-bamm input
Command line program with many options.
Mandatory options -d design file -m model
file -t template image -o output directory
BAMM
24
x-bamm input
Design Matrix Lists all files in the analysis
plus independent variable and covariates /home/us
er/images/case1/case_1.004.GST1 1
45 /home/user/images/case1/case_2.005.GST1 1
44 /home/user/images/case1/case_7.002.GST1 0
32 ...
BAMM
25
x-bamm input
Model description Indicates which values of the
design matrix to use, which is the independent
variable, which are the covariates. Example 2x2
factorial design MAIN 1 2 INTERACTIONS 1 2
BAMM
26
x-bamm input
Necessary options -l design u1x1u2x1p2x1u3x1p
3x1u2x2p2x2 -c which column is the independent
variable (0 overall for 3x1 designs) -f data type
(fmristruct) -p p-value (auto give ecpilt1)
BAMM
27
x-bamm input
Optional options -D cluster dimension (23) -v
Add the total activation/tissue as a covariate -g
2D smoothing
BAMM
28
x-bamm input
Leave itits not worth it. -C cluster
threshold -L logs of dependent variable -b
background threshold -s standardisation of test
statistic -a cluster statistic
BAMM
29
x-bamm output
Images ClusterTestp0.???tal.ppm shows the
effects specified by the -c variable. Colours
(red/yellow blue/purple) denote relative
increase/decrease in effect. Depends on effect
and class codings. ClusterTestp0.???tal.info
also.
BAMM
30
x-bamm output
-----Cluster Significance Testing----- Randomised
Data Range -750.725927 -gt
478.872661 No. of Clusters(-ve) 1066 No. of
Clusters(ve) 997 Total
2063 CV(-ve) -38.898803 CV(ve)
36.794414 Observed
Data Range -51.350409 -gt
295.429191 No. of Clusters(-ve) 59 No. of
Clusters(ve) 166 Total
225 Significant Clusters(-ve)1 Significant
Clusters(ve)10 Total
11 Estimated No. of FPs lt11 Confidence
intervals /-6.4 -------------------------------
-----
Text file bamm.log Contains runtime and stats
information (including errors)
BAMM
31
x-bamm output
ClusterTestp0.???.dat 166.4 234.1 506.3 ...
-1.0 345.2 211.7 112.3 ... 1.0
-1
1
ClusterTestp0.???tal.ppm
BAMM
32
I am pleased to inform you that your submission
has been accepted for publication without
revision.
?
BAMM