Hello all,
Need assistance in building a design for LONGITUDINAL DATA IN TWO GROUPS WITH INTERACTION OF COVARIATE. I am referring to the CAT12 manual page 45 of 12.8 (1830).
I have two groups (experimental and control) and each has 4 time points. I have the GM maps. I want to see the correlation between the behaviour score in the experimental group. I learnt that I can put the scores as covariates and check the regions affected by the scores for one group as well, as in the manual it is written:
"This model supports the analysis of linear increase or decreases over time that differ between two groups.
In order to analyse non-linear time effects, a polynomial covariate can be used. Furthermore, this model
can be also reduced to one group accordingly or you can use other parameters such as performance over
time to analyse dependencies to that variable (e.g. linear increase w.r.t. to performance)."
However, when I just input scans for experimental group,
First method
with two factors (Subjects and Time points)
-> under Scans: smooth maps of 5 time points
-> under conditions: [1 2 3 4 ] (four columns for time points)
-> covariates -> score1 ->interactions: with factor 2( which is time point)
Second method
with three factors (Subjects, Groups, Time points)
-> under Scans: smooth maps of 4 time points
-> under conditions: [1 1 1 1, 1 1 1 1 ] (two columns, one for group 1 and second column with all time points as 1 as it was mentioned in the manual: all time points are coded here with only one value. The
differentiation between the time points is now made by the covariate. )
-> covariates -> score1 ->interactions: with factor 2 ( which is group)
I am unable to figure out how shall I design the model in order to carry out correlation between score and longitudinal data in one group?
Need assistance.
Thank you
Regards,
Jhelum
|