Hi everyone,

Quick question. I currently have multiple symmetric 68*68 AEC-c matrices (each cell in the matrix contains an AEC-c measure between two locations in the brain on source level). I've got two measurement moments (baseline/follow-up), two different meds (Aripiprazole or Amisulpride) and five different brain frequencies I want to look at. At the moment I've got a total of 2*2*5 = 20 matrices of 68*68. 

The thing I want to compare is differences in connectivity between the two measuring moments while also looking at more detailed differences within the brain frequencies and potential differences between meds.

Should I just make a matrix with all the AEC-c measures in one row while inserting extra row names?

Eg:

Row 1: Measurement A | Aripiprazole | Theta brain wave | AEC-c measurements (2312 (68*68/2) columns)
Row 2: Measurement C | Amisulpride | Delta brain wave | AEC-c measurements (2312 (68*68/2) columns)

Hi Suzannah,

I don't fully follow your experimental design, but here are a few quick considerations: 

- You might want to start by analyzing each frequency independently, rather than testing for interaction effects across frequencies. 

- If the two measurements are made in the same subjects, you will need a repeated measures design. Check out the manual or this forum for examples of a repeated measures design. 

I imagine that your design matrix would include a column for time (baseline/follow-up), time x medication interaction and columns to model the mean for each subject. 

best wishes and happy to answer any specific questions,
Andrew

Originally posted by Suzannah Cunnington: