Premium
IC‐P‐167: Detection of changes in functional connectivity following aricept ® treatment in mild Alzheimer disease subjects
Author(s) -
Li ShiJiang,
Ward Douglas B.,
Wu Zhilin,
Jones Jennifer,
McRae Thomas,
Franczak Margorzata,
Antuono Piero
Publication year - 2009
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.2009.05.140
Subject(s) - mathematics , wilcoxon signed rank test , pattern recognition (psychology) , nonparametric statistics , region of interest , artificial intelligence , resting state fmri , statistics , nuclear medicine , medicine , computer science , mann–whitney u test , radiology
Background: Resting state fMRI has been used to measure functional connectivity (fcMRI) in discrete brain areas. We have now developed a global fc analysis on a large number of anatomically predefined ROIs to determine if it could separate AD from controls and to observe if three-month treatment with donepezil (Aricept ) could improve the functional connectivity of the neuronal network. Methods: Twenty AD and 20 control subjects, as well as 14 mild AD subjects (MMSE 23-27) who received 3 months of Aricept administration, were recruited for the study. The high-resolution anatomical image for each subject was transformed and aligned with a reference template containing 116 anatomically defined ROIs. The average time course within each ROI was extracted from the functional imaging data, and then detrended, filtered, and motion corrected. Functional connectivity between ROI pairs was assessed by calculating the Pearson product-moment correlation coefficient (cc). For each pair of ROIs, the nonparametric two-sample Wilcoxon rank-sum test was performed and a matrix of z-values was obtained (referred to as the ‘‘Wmatrix’’). The histogram of these W-matrix z-values was used to select two subsets of ROI pairs according to high or low cc values. The accuracy of the classification algorithm was evaluated using the unbiased ‘‘leave-one-out’’ (LOO) method. Results: With the Fisher Linear Discriminant test, all of the Controls are located above the linear boundary (Blue Network), and all but one of the AD study subjects are located below the linear boundary (Red Network). The area under the ROC curve is 0.88. A three-month treatment with donepezil demonstrated an improvement (p < 0.013) in the global functional connectivity from baseline. Conclusions: The global ROI-based resting-state fcMRI analysis allows discrimination between AD and Control populations with high sensitivity and specificity. Functional connectivity was improved after three months of treatment with Aricept suggesting an important biological role of the compound on neuronal networks and that fcMRI method is a promising methodology to monitoring neural responses to therapeutic treatment.