Faculty Collaboration Database

COmplex-Model-Based Estimation of thermal noise for fMRI data in the presence of artifacts. Magn Reson Imaging 2007 Sep;25(7):1079-88

Date

08/21/2007

Pubmed ID

17707170

Pubmed Central ID

PMC2699626

DOI

10.1016/j.mri.2006.12.007

Scopus ID

2-s2.0-34547782158 (requires institutional sign-in at Scopus site)   6 Citations

Abstract

Due to the presence of artifacts induced by fast-imaging acquisition in functional magnetic resonance imaging (fMRI) studies, it is very difficult to estimate the variance of thermal noise by traditional methods in magnitude images. Moreover, the existence of incidental phase fluctuations impairs the validity of currently available solutions based on complex datasets. In this article, a time-domain model is proposed to generalize the analysis of complex datasets for nonbrain regions by incorporating artifacts and phase fluctuations. Based on this model, a novel estimation schema has been developed to find an appropriate set of voxels in nonbrain regions according to their levels of artifact and phase fluctuation. In addition, noise intensity from these voxels is estimated. The whole schema is named COmplex-Model-Based Estimation (COMBE). Theoretical and experimental results demonstrate that the proposed COMBE method provides a better estimation of thermal noise in fMRI studies compared with previously proposed methods and suggest that the new method can adapt to a broader range of applications, such as functional connectivity studies, evaluation of sequence designs and reconstruction schemas.

Author List

Xu Y, Wu G, Rowe DB, Ma Y, Zhang R, Xu G, Li SJ

MESH terms used to index this publication - Major topics in bold

Algorithms
Artifacts
Computer Simulation
Hot Temperature
Humans
Image Processing, Computer-Assisted
Magnetic Resonance Imaging
Monte Carlo Method
Statistics as Topic