[[FsgdExamples|Back to FSGD Examples]]

~+'''Six Groups (Two Factors/Three Levels), No Covariates'''+~

This models the input as six separate intercepts/means/offsets, one
for each group. The six groups groups can be thought of as three
levels of a two discrete factors (eg, Gender (Male, Female),
Handedness (Left, Right, Ambidextrous)). There are not continuous
covariates.  These results hold only for DODS or DOSS.

<<TableOfContents>>

= FSGD File (g6v0.fsgd) =
{{{
GroupDescriptorFile 1
Title OSGM
Class MaleRight
Class MaleLeft
Class MaleAmbi
Class FemaleRight
Class FemaleLeft
Class FemaleAmbi
subject1 MaleRight
subject2 MaleLeft
subject3 MaleAmbi
subject4 FemaleRight
subject5 FemaleLeft
subject6 FemaleAmbi
}}}

Nclasses = 6 <<BR>>
Nvariables = 0 <<BR>>

= Regressors =

NregressorsDODS = Nclasses*(Nvariables+1) = 6*(0+1) = 6 <<BR>>
NregressorsDOSS = Nclasses + Nvariables = 6 + 0 = 6 <<BR>>
Regressor1: ones for MaleRight subjects, 0 otherwise. Codes intercept for Group 1 <<BR>>
Regressor2: ones for MaleLeft subjects, 0 otherwise. Codes intercept for Group 2 <<BR>>
Regressor3: ones for MaleAmbi subjects, 0 otherwise. Codes intercept for Group 3 <<BR>>
Regressor4: ones for FemaleRight subjects, 0 otherwise. Codes intercept for Group 4 <<BR>>
Regressor5: ones for FemaleLeft subjects, 0 otherwise. Codes intercept for Group 5 <<BR>>
Regressor6: ones for FemaleAmbi subjects, 0 otherwise. Codes intercept for Group 6 <<BR>>

= Contrasts =

The number of columns in each contrast matrix must be the same as the
number of regressors (Nregressors). If there is only one row in the
contrast matrix, then the result will be a t-test and will have a
sign. Reversing the signs in the contrast matrix will only change the
sign of the output, not its magnitude.  If there is more than one row,
the result will be an F-test and will be unsigned.

== Contrast 1 male.right-vs-left.mtx ==

Null Hypothesis: is there a difference between the Male Righthanders
and Male Lefthanders?

{{{
1 -1 0 0 0 0
}}}

This is a t-test with MR>ML being positive (red/yellow).

== Contrast 2 male.right-vs-ambi.mtx ==

Null Hypothesis: is there a difference between the Male Righthanders
and Male Ambidextrous?

{{{
1 0 -1 0 0 0
}}}

This is a t-test with MR>ML being positive (red/yellow).

== Contrast 3 female.right-vs-ambi.mtx ==

Null Hypothesis: is there a difference between the Female Righthanders
and Female Ambidextrous?

{{{
0 0 0 1 0 -1
}}}

This is a t-test with MR>ML being positive (red/yellow).

== Contrast 4 male-female.mtx ==

Null Hypothesis: is there a difference between Males and Females
regressing out the effects of handedness?

{{{
[1 1 1 -1 -1 -1]/3
}}}

This is a t-test with M>F being positive (red/yellow).  This could
also have simply been [1 1 1 -1 -1 -1] without effecting the
significance. But using 1/3 makes the gamma.mgh output
volume have the same scale. Eg, if the input is thickness in mm, then
the output will be the mean MF difference in thickness also meausred
in mm.

== Contrast 5 right-left.mtx ==

Null Hypothesis: is there a difference between Righties and Lefties regressing out the effects of gender?

{{{
[1 -1 0 1 -1 0]/2
}}}

This is a t-test with R>L being positive (red/yellow).See Contrast 4 for notes on scale.

== Contrast 6 gender-x-handedness.mtx ==

Null Hypothesis: is there an interaction between gender and handedness?

{{{
|+1 -1  0 -1 +1  0|
|+1  0 -1 -1  0 +1|/2
}}}

This contrast matrix has two rows, so this will be tested with an F-test.

= mri_glmfit command =

This is an example invocation of mri_glmfit. Depending upon your application, you may have other options as well.

{{{
mri_glmfit \
  --glmdir g6v0 \
  --y y.mgh \
  --fsgd g6v0.fsgd \
  --C male.right-vs-left.mtx \
  --C male.right-vs-ambi.mtx \
  --C female.right-vs-ambi.mtx \
  --C male-female.mtx \
  --C right-left.mtx \
  --C gender-x-handedness.mtx 
}}}