LME Matlab tools. Author: Jorge Luis Bernal Rusiel, 2012. jbernal@nmr.mgh.harvard.edu or jbernal0019@yahoo.es

This directory contains longitudinal data that can be used to get started with the LME tools. There are two types of data: univariate responses (univariate subdirectory) and mass-univariate spatial data (mass_univariate subdirectory). 

The univariate directory contains four archives: cities6.mat (Six Cities Study of Air Pollution and Health), menarche.mat (Influence of Menarche on Changes in Body Fat Accretion), aid_cd4.mat (Randomized Study of Dual or Triple Combinations of HIV-1 Reverse Transcriptase Inhibitors) and ADNI791_Hipp_and_Entorh.mat. The first three archives contain data for three famous longitudinal studies. These data were downloaded from http://biosun1.harvard.edu/~fitzmaur/ala/  and detailed analyses of them can be found in Chapter 8 of the book: Fitzmaurice, G.M., Laird, N.M., Ware, J.H., 2004. Applied longitudinal analysis. Wiley.

The  ADNI791_Hipp_and_Entorh.mat archive contains longitudinal data for a sample of 791 subjects of the Alzheimer Disease Neuroimaging Initiative (ADNI) cohort. These subjects belong to four different groups: (1) Stable healthy control (the reference group): those who were clinically healthy throughout the follow-up period. (2): Stable MCI (mci_stable): those who were categorized MCI at baseline and remained so throughout the study. (3) Converter MCI (mci_conv): those who were MCI at baseline and progressed to the dementia phase of AD during follow-up. (4) AD patients (alzh): those who were diagnosed with dementia of the Alzheimer type at baseline.  Two biomarkers were measured, namely mean thickness within the Entorhinal cortex and Hippocampal volume, since these are two classical MRI-derived markers that are known to be strongly associated with early AD. The data were processed with FreeSurfer's longitudinal stream (v5.1).

The mass_univariate subdirectory contains spatial longitudinal data for a random subsample of the ADNI cohort of 50 Stable MCI (the reference group) and 50 Converter MCI (cMCI) subjects. There are three archives ADNI_Long_50sMCI_vs_50cMCI.mat, lh.50sMCI_vs_50cMCI_long_thickness_sm10.mgz and rh.50sMCI_vs_50cMCI_long_thickness_sm10.mgz. The first archive contains the design matrix and the surfaces and cortex label (erode by two steps) of the average template subject (fsaverage). The other two are Freesurfer's .mgz archives containing cortical thickness data for corresponding brain hemispheres. These data were smoothed with an iterative nearest neighbour averaging procedure that approximates a 10mm Gaussian kernel smoothing along the surface. You can read these .mgz files into matlab using the fs_read_Y function provided with the toolbox. The cortical thickness data were also processed with FreeSurfer's longitudinal stream (v5.1).
