JSM 2012 Home

JSM 2012 Online Program

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

Online Program Home

Abstract Details

Activity Number: 177
Type: Contributed
Date/Time: Monday, July 30, 2012 : 10:30 AM to 12:20 PM
Sponsor: Biometrics Section
Abstract - #306829
Title: A New Nonparametric, Information-Splitting Image Analysis Technique
Author(s): Mark Inlow*+ and Jing Wan and Sungeun Kim and Kwansik Nho and Shannon Risacher and Andrew Saykin and Li Shen
Companies: Rose-Hulman Institute of Technology and Indiana University School of Medicine and Indiana University School of Medicine and Indiana University School of Medicine and Indiana University School of Medicine and Indiana University School of Medicine and Indiana University School of Medicine
Address: 2117 N 10th St, Terre Haute, IN, 47804-2306, United States
Keywords: neuroimaging ; image analysis ; random field theory ; left-spherical distribution ; nonparametric
Abstract:

We present a nonparametric alternative to random field theory (RFT) image analysis methods. Our procedure uses a new information splitting approach based on left-spherical distribution theory; it splits the sample information by assigning two independent statistics, a chi-square and a t, to each voxel. The stochastic independence of these two sets of statistics (statistical parametric maps) follows by left-spherical distribution theory if the general linear model statistical parametric map is multivariate normal. The chi-square statistics are used to identify non-null voxels allowing enhanced analysis of the remaining information provided by the t statistics. There is no data snooping penalty since the chi-square and t statistics are independent. P-values are computed via permutation to avoid additional assumptions and to provide robustness against violation of the multivariate normal assumption. We propose using test statistics with normal permutation distributions to allow t-tests based on small numbers of permutations. We compare the performance of our approach with that of RFT (SurfStat) on hippocampal data.


The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2012 program




2012 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.