Premium
IC‐P‐083: A JOINT SURVIVAL‐LONGITUDINAL MODELLING FOR DYNAMIC PREDICTION OF PROGRESSION FROM MCI TO AD IN ADNI STUDY
Author(s) -
Abderazzak Mouiha,
Duchesne Simon
Publication year - 2014
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2014.05.088
Subject(s) - biomarker , medicine , time point , random effects model , mixed model , oncology , nuclear medicine , statistics , mathematics , biology , physics , biochemistry , meta analysis , acoustics
aligned to the corresponding baseline pair using the proposed multi-modal registration technique. The resulting longitudinal maps were normalized to an inter-subject average for comparison. Voxel-wise statistical analysis was performed using permutation test (n1⁄45000) with age and gender as covariate. For comparison, we performed the same experiment using only the T1w scans.Results: Figure 1-bottom shows areas of significant difference in rate of volume change. It can be noticed that using the proposed multi-modal method, significant areas in thewhite matter are removed.We argue that volume change differences in the white matter obtained from T1w MRI only are a consequence of the registration regularisation and are thus erroneous findings. Conclusions: We present a morphometric analysis based on multi-modal image registration. This technique could provide data driven findings that are less susceptible to error introduced by algorithmic choices of the registration. All software used to perform this analysis has been made open-source to enable others to conduct similar experiments.