Modeling and Analysis of Inherently Complex-Valued FMRI Data
Daniel
Rowe

Marquette University

Thursday, April 14, 2016 - 3:30pm

In fMRI and fcMRI, the original data measured by the MRI machine are complex-valued, a complex-valued inverse Fourier transform is applied to reconstruct into a complex-valued image. Almost exclusively, the Cartesian real and imaginary images are converted to magnitude and phase images, then the phase half of the data is discarded before statistical analysis. A description of potential biological information in the phase will be provided along with models to compute brain activation using both magnitude and phase. Results from both simulated and experimental data will be presented.