# Thread: Repeated Measures model with Covariates

1. ## Repeated Measures model with Covariates

Hi Everyone, i am struggling with finding a good way to analyze a dataset. The design is this: I have 25(n) subjects, all of them are tested in an experiment under 9 conditions. I estimate an electrophysiological index (idx) in every condition for each subject. Now I want to ask whether the volume of brain structures a,b,c has an effect on this index(idx) and whether this differs between conditions.
What I think is that this is a repeated measures model where i have for every subject the output of the 9 conditions (9 *idx), and for every of them I have a value for brain structure a,b,c. The model I am trying to fit is:
y1-y9 ~ a,b,c
where (y1-y9) = indices in the nine conditions for every subject. a,b,c = volume of structure a,b,c for every subject.
I am not sure though this is the right way to do this and also I am not sure how to interpret the results. I specify a within subject model, is this right? Another possibility could be using a linear model where:
idx ~ (a+b+c)*factor1*factor2 where idx=column vector of all indices , and factor 1 and 2 are categorical values (covariates) specifying the levels of factors for every row (i.e. idx ). a,b,c are repeated (since they are the same across conditions).

2. ## Re: Repeated Measures model with Covariates

Hi Cecilia, as far as I see you can use a mixed regression model (or mixed ANOVA) of this type:

idx ~ brain_structure + condition + brain_structure:condition,

with "subject_ID" as a random factor. Thus, the outcome related to "brain_structure" tells you if brain structure effects the index, analogous for "condition" and the interaction term "brain_structure:condition" tells you if the dependency of the index on brain_structure differs under different conditions respectively the dependency of the index on condition changes with different brain structures.

 Tweet