I'm going to suppose you did the pipeline thing over your MRI images and you already have the study_mri.csv file. The Freesurfer version I'm using makes an estimation of Intracraneal volume and the pipeline scripts put it as EstimatedTotalIntraCranialVol_eTIV into the results file. First do this for a real CSV file:

$ sed 's/;/,/g' ab255_mri.csv > ab255_mri_rc.csv

Let's say we want to get the hippocampus volume (HV) corrected by ICV. Then we must get Left and Right HV and calculate its medium value. HV = 0.5*(HVR+HVL)

Now first step is to fit a straight line for HV vs ICV and the ICV mean value (<ICV>). The adjusted HV will be aHV = HV - b*(ICV-<ICV>), with b as the regression coefficient of HV vs ICV.

Now we need to get the data. Let's try it in R.

x <- read.csv("ab255_mri_rc.csv")
x$HV = 0.5*(x$"Right.Hippocampus"+x$"Left.Hippocampus")
a <- lm(x$HV~x$"EstimatedTotalIntraCranialVol_eTIV")
b=a$coefficients[[2]]
x$aHV = x$HV - b*(x$EstimatedTotalIntraCranialVol_eTIV - mean(x$EstimatedTotalIntraCranialVol_eTIV, na.rm=TRUE))
write.csv(x, file="ab255_mri_aHV.csv")

El paquete ADNIMERGE provee los valores de Hipoocampo (suma de ambos) e ICV. Puede corregirse directamente y dividirse entre 2 para obtener valores similares.

library("ADNIMERGE")
a <- lm(adnimerge$Hippocampus~adnimerge$ICV)
adnimerge$adj.Hippocampus = 0.5*(adnimerge$Hippocampus - a$coefficients[[2]]*(adnimerge$ICV - mean(adnimerge$ICV, na.rm=T)))

Este procedimiento se debera aplicar para realizar la replica de cualquier estudio donde se utilize la semisuma corregida de ambos hipocampos.