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

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