Dear SPM people,
I have exhaustively looked for an answer to this before posting. I
definitely need your kind help.
In a second level analysis, I need to explore the variables I am playing
with, but I cannot do that.
The plot is as follows:
I have a contrast with three separate conditions, an F contrast: [1 0 0;
0 1 0; 0 0 1]’.
These first three columns represent three groups that I have introduced
in an ANOVA. The rest of the columns are covariates.
I would like to know the response in any group, so I ask the
aforementioned F contrast.
I get an FWE p<.05 region. I go to the global maxima and plot the estimates.
They look like the image I attach. Estimation of activation is nearly 0
in one group while greater than zero in the other two groups.
Now it is that I get into problems: I get these nice parameter
estimation. Therefore, I would like to look into my data and see whether
there are big differences between group one (i.e. bar one) and the rest
of them. I would like to see how the parameters affect my data. I expect
to have two groups of activation, one near to zero and the other two
greater than zero.
I ask for a plot of 'fitted responses'. Then, if I ask for 'predicted'
results then I obtain Y with the betas for each contrast, not the
estimated signal.
If I otherwise ask for 'adjusted', then I get a set of heavily
overlapped data (see attached image).
Any suggestion as to how to proceed would be greatly appreciated!
Thanks in advance,
Juanjo
--
________________________________________________________
Juan J. Lull Noguera - jualulno_at_upvnet.upv.es
- [log in to unmask]
http://www.ibime.upv.es
[MI - Medical Imaging Area]
IBIME::Biomedical Informatics Group
UPV - Politechnical University of Valencia - Spain
________________________________________________________
[IM - Área de Imagen]
IBIME::Informática Biomédica
UPV - Universidad Politécnica de Valencia
________________________________________________________
"Education is the best provision for the journey to old age"
Aristotle
|