Dear Free-surfer Group, I am currently attempting to use the lme matlab tools for my longitudinal analysis. My dataset consists of two time-points, and I am interested in the group by time interaction for my dataset.
I am attempting to follow the wiki instructions, but are finding them hard at times. I have reached the stage where, for the mass-univariate analysis", it says to do
[lhTh0,lhRe] = lme_mass_fit_EMinit(X,[1 2],Y,ni,lhcortex,3);
May I just ask if its correct to fit the model with two random effects when I only have two time-points. Furthermore, this command is currently running, and it keeps saying, over and over,
Warning: Matrix is close to singular or badly scaled. Results may be inaccurate. RCOND = 7.100165e-19.
In lme_mass_fit_EMinit>(parfor body) at 98
In parallel_function>make_general_channel/channel_general at 907 In remoteParallelFunction at 28
Has anybody else recieved this error? Is there a simpler method whereby to fit my model and the matrix, ie is there actually something wrong with my design matrix? It is as follows:
intercept time hr+ 3xtime hr- 5xtime hc 7xtime hrill 9xtime age
First column, all ones (intercept)
Second column; time
third column; group1
fourth column; group1*time
fifth column; group 2
sixth column; group2*time
seventh column; group three
eighth column; group three*time
nineth column; group4
tenth column; group4*time
eleventh column; age
Does this look correct? Can anyone please advise me on the error I am getting?
Cheers,
Melinda