How to code a repeated measure variable that is the independent variable for a mixed effect model

Hello,

I am a total beginner. I am trying to run a mixed effect model with memory performance (LDI) as a dependent variable, and a functional measure that was taken from the same brain region in the left and in the right hemisphere as an independent variable (CVR HIP). In addition I would like to control for age, sex and site (data was collected at 2 different scanning sites).

The problem I have is that I seem to not be able to figure out how to code the 2 brain measures as one independent variable with a left and right variable in order to include it into the model.

This is how I coded the data:

and then I ran:

fit1=fitlme(data, 'LDI~CVRHIP+hemisphere+age+sex+site+(1|subject)')

But I am pretty sure that this is not correct. When I look at the fixed effects coefficients table I believe the effect 'hemisphere_2' gives me the effect for the hemisphere 2 on memory performance in comparison to hemisphere 1 but I want to look at the effect of CVR in hemisphere 2 on memory in comparison to effect of CVR in hemisphere 1.

I hope I explained everything in an understandable way. I would very much appreciate any comments or suggestions. Thank you very much in advance!

Carolin

Answer 1

[Hitesh]

Hi Caroline chwiesko,

You can get effect of "CVR_HIP" in "hemisphere" 2 in comparison to effect of "CVR_HIP" in "hemisphere" 1 by using below "fitlme" expression'. You can also remove constant variable such as "site" which is always 1 as it doesn't contribute any useful information.

fit1 = fitlme(data, 'LDI ~ CVR_HIP * hemisphere + age + sex + (1|subject)');

This model includes both the main effects of "CVR_HIP" and "hemisphere" as well as their interaction "CVR_HIP * hemisphere". This means that the model can estimate how the effect of "CVR_HIP" on "LDI" differs between hemisphere 1 and hemisphere 2. "CVR_HIP:hemisphere" represents how the effect of "CVR_HIP" on "LDI" changes when moving from hemisphere 1 to hemisphere 2.

For more information on "fitlme", refer to the below MATLAB documentation:

https://www.mathworks.com/help/stats/fitlme.html#:~:text=the%20morning%20shift.-,LME%20Model%20for%20Split%2DPlot%20Experiment,-Longitudinal%20Study%20with