Partially Adaptive STAP for fMRI: A Method for Detecting Brain Activation Regions in Simulation and Human Data

Citation data:

CONFERENCE: Biomedical Imaging: From Nano to Macro, 2007

Biomedical Imaging: From Nano to Macro, 2007, Vol: 0

Publication Year:
Usage 39
Abstract Views 39
Repository URL:
Thompson, Elizabeth A.; Huang, Lejian; Holland, Scott K.; Schmithorst, Vincent; Talavage, Thomas M.
Institute of Electrical and Electronics Engineers
biomedical; MRI; brain; medical signal processing; thompson; Engineering
lecture / presentation description
This paper introduces three partially adaptive space-time processing (STAP) schemes for analyzing fMRI data. Element space partially adaptive STAP can achieve performance close to that of fully adaptive STAP while greatly decreasing the CPU running time and memory requirements when applied to both synthetic as well as real human brain data. In synthetic analyses, partially adaptive STAP algorithms exhibit better detection characteristics than the traditional cross-correlation method. This is supported by human data in which element space and fully adaptive STAP produce activation maps that closely resemble those of cross-correlation.