|
Activity Number:
|
463
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #309749 |
|
Title:
|
Detecting Activations in fMRI Experiments with Maximum Cross-Correlation Statistics
|
|
Author(s):
|
Kinfemichael Gedif*+ and William R. Schucany and Richard Gunst and Wayne Woodward and Jeffrey Spence and Patrick Carmack and Qihua Lin
|
|
Companies:
|
Southern Methodist University and Southern Methodist University and Southern Methodist University and Southern Methodist University and The University of Texas Southwestern Medical Center at Dallas and The University of Texas Southwestern Medical Center at Dallas and The University of Texas Southwestern Medical Center at Dallas
|
|
Address:
|
6236 Ridgecrest Rd Apt 2424, Dallas, TX, 75231,
|
|
Keywords:
|
fMRI ; cross-correlation ; HRF ; kernel-smoothing
|
|
Abstract:
|
Cross-correlation between an anticipated hemodynamic response function (HRF) and the observed voxel's time course is one method of identifying activated voxels due to the known stimulus sequence in an fMRI experiment. In fact, such a method is an optimal detector if the underlining noise process is Gaussian and the true signal of interest is known. In practice however, the true signal of interest is not known and analysis based on fitting HRF models strongly depends on the adequacy of the fitted model. We propose an approach that does not require fitting an HRF to the voxel time series. The maximum cross-correlation between the kernel-smoothed ideal stimulus sequence and shifted (lagged) values of the observed response is the proposed test statistic. Improved efficiency is demonstrated in simulations involving realistic drift and noise models.
|
