I'm trying to perform VBM analysis using FSL.
I've successfully set up everything for my vbm analysis using FSL,
such as placing all the files in the same directory and re-scaling
the intensity values.
But at am stuck at setting up the design and contrast matrices.
I have 3 groups named control, short and long. Each group has 9
I've been using the GLM setup GUI and have selected the higher level/
non-time-series analysis and told it there are 27 inputs.
I've assigned the first 9 inputs to group 1, the second 9 inputs to
group 2 and the final 9 to group number 3.
Am I correct that I need 3 EVs for this model? Do I then need to set,
for example, EV1 to 1 only for the 9 inputs in group 1 and likewise
for the other two groups?
Under the EV1, EV2 and EV3 labels in the GLM GUI there are text
boxes, what are these for?
The contrasts and F-tests tab defaults to "group mean" which I guess
is ok for functional analysis but not for a VBM analysis. Is there
even an equivalent for VBM analysis?
Are appropriate contrasts for this analysis control - short, control
- long and should I do F tests for each of these contrasts?
When I have 3 groups such as in this case do I need to use the -g
option to randomise when performing permutation testing?
Are the inputs read in the order in which they are concatenated in
the 4D file? So if fslvbm_3_proc concatenates in alphabetical order
will the first 9 inputs correspond to the first 9 volumes in the 4D
file which in this case will be the 9 control subjects? I'm assuming
that the next 9 volumes will be the 9 long subjects and the final 9
volumes the 9 short subjects. Does the order of the subjects within
group matter? So does EV1 correspond to the control group, EV2 to the
long and EV3 to the short groups?
I noticed that the fslvbm_1_bet defaults to passing -f 0.4 to BET. If
I run "fslvbm_1_bet -N -B -f 0.3" which option will bet use the 0.4
Thanks in advance,
Dr Colm G. Connolly
School of Psychology and Institute of Neuroscience
The Lloyd Building
University of Dublin
Trinity College, Dublin 2, Éire