1
$\begingroup$

I have 139 subjects (ID), with measurements taken at two time points (Time1, Time2), at 148 brain regions, a dependent measure called volume, and a covariate called thickness.

Each subject has 148 brain regions with volume and thickness measured twice

I am trying to find out if there is a difference in volume between timepoint 1 and timepoint 2 while controlling for thickness. I want to know which brain regions show this difference. I need help setting up the model. Specifically the timepoint part is throwing me off...

I am using R. and trying to figure out a model with linear mixed models with (1|ID) as random factor, fixed factors regions, thickness.

I was thinking lmer(volume ~ thickness + (1 | ID / regions)?

EDIT: lmer(volume ~ thickness + timepoint + (1 | ID / regions)`

Linear mixed model fit by REML. t-tests use Satterthwaite's method ['lmerModLmerTest'] Formula: volume ~ thickness + timepoint + (1 | ID/regions) Data: DATA

REML criterion at convergence: -1704.6

Scaled residuals: Min 1Q Median 3Q Max -6.5771 -0.2711 -0.0559 0.1816 9.6790

Random effects: Groups Name Variance Std.Dev. regions:ID (Intercept) 0.06566 0.2562
ID (Intercept) 0.01917 0.1385
Residual 0.01506 0.1227

Fixed effects: Estimate Std. Error df t value Pr(>|t|)
(Intercept) 9.247e-02 3.533e-02 8.500e+01 2.617 0.0105
thickness 1.449e-01 9.615e-03 7.607e+03 15.068 <2e-16

timepoint1 -1.320e-02 1.349e-03 4.086e+03 -9.787 <2e-16

Correlation of Fixed Effects: (Intr) thickness thickness -0.661
timepoint1 0.017 -0.026

  1. What is the intercept for fixed effects?
  2. How can I answer if there was a significant increase or decrease in volume from time point 1 to timepoint 2?
  3. Can I obtain regional effects? i.e. Region 12 increased from timepoint 1 to time point 2 ? Proposed Model:
    MODEL2 = lmer(volume~ thick + timepoint + regions + (1|ID/regions), data = DATA )
$\endgroup$
2
$\begingroup$

Since you are interested in differences in volume between the two time points, you would need to include the time variable as a fixed effect as well, i.e.,

lmer(volume ~ time + thickness + (1 | ID / regions))

where time is is binary variable taking the value 0 for the first time point, and 1 for the second one.

$\endgroup$
  • $\begingroup$ Thank you for your reply! I have time as a factor 1, 2. Is it necessary to make it binary 0,1 ? data.frame': 8288 obs. of 5 variables: $ volume : num 579 951 229 286 844 ... $ thick : num 2.37 2.09 1.94 2.6 2.78 ... $ timepoint: Factor w/ 2 levels "1","2": 1 1 1 1 1 1 1 1 1 1 ... $ regions : Factor w/ 148 levels "1","2","3","4",..: 1 2 3 4 5 6 7 8 9 10 ... $ ID : Factor w/ 139 levels "1","2","3","4",..: 1 1 1 1 1 1 1 1 1 1 ... $\endgroup$ – Sheraz Jan 6 at 21:24
  • $\begingroup$ No, if you have it as factor it will also work. $\endgroup$ – Dimitris Rizopoulos Jan 6 at 21:25
  • $\begingroup$ I edited my post with follow up. I am confused what the fixed effects intercept means here? Added some questions. Thank you so much for your help $\endgroup$ – Sheraz Jan 6 at 21:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.