|
Activity Number:
|
101
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #304986 |
|
Title:
|
Multiscale Adaptive GEE Methods for Longitudinal Imaging Data
|
|
Author(s):
|
Yimei Li*+ and Hongtu Zhu and Joseph G. Ibrahim and Dinggang Shen
|
|
Companies:
|
The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill
|
|
Address:
|
Biostatistics Department, Chapel Hill, NC, 27599,
|
|
Keywords:
|
Asymptotic Properties ; Multiscale adaptive regression ; GEE ; Wald statistics ; Voxel-wise method ; Weights
|
|
Abstract:
|
We develop a multiscale adaptive GEE method (MARMGEE) for spatial and adaptive analysis of longitudinal imaging data. The primary motivation and application of the proposed methodology is statistical analysis of imaging data on the two-dimensional (2D) surface or in the 3D volume for neuroimaging studies. The key idea of the MARMGEE is to successively increase the radius of a spherical neighborhood around each voxel and combine all the data in a given radius of each voxel with appropriate weights to adaptively calculate parameter estimates by GEE method and test statistics. We establish consistency and asymptotic normality of the adaptive GEE estimates and the asymptotic distributions of the adaptive test statistics. Particularly, we show theoretically that the MARMGEE outperforms classical voxel-wise GEE approach.
|
