The comparison of different methods for MCI-to-AD conversion prediction using ADNI data is hampered by the
fact that the nearly all works use a different classification of the subjects into stable (non-converters) and
progressive (converters) MCI. To simplify the comparisons, we provide a mat-file specifying the labeling used in our paper:
Moradi, Elaheh, Pepe, Antonietta, Gaser, Christian, Huttunen, Heikki, Tohka, Jussi, ADNI. Machine learning framework for early MRI-
based Alzheimer’s conversion prediction in MCI subjects, NeuroImage (2014), doi:10.1016/j.neuroimage.2014.10.002
The conversion data was downloaded in April 2014 and the criteria to define MCI groups are explained in the paper.
Briefly, we defined subjects as
(1) sMCI (stable MCI), if diagnosis was MCI at all available time points (0-96 months), but at least for 36 months (n = 100);
(2) pMCI (progressive MCI), if diagnosis was MCI at baseline but conversion to AD was reported after baseline within 1, 2 or 3 years, and without reversion to MCI or NC at any available follow-up (0-96 months) (n= 164)
(3) uMCI (unknown MCI), if diagnosis was MCI at baseline but the subjects were missing a diagnosis at 36 months from the baseline or the diagnosis was not stable at all available time points (n = 100).
[Download the mat-file] [Download preprint]
The description of the file: The first column of the MCI matrix gives subjects' RIDs. The second column gives the label in the following coding:
labels:
1: pMCI converted to AD within the first 12 months
2: pMCI converted to AD between 12-24 months
3: pMCI converted to AD between 24-36 months
4: sMCI
5: uMCI
The same information can be found in the supplement to the paper.