Complex denoising of MR data via wavelet analysis: Application for functional MRI

A fast post-processing method for noise reduction of MR images, termed comp lex-denoising, is presented. The method is based on shrinking noisy discret e wavelet transform coefficients via thresholding, and it can be used for a ny MRI data-set with no need for high power computers. Unlike previous wave let application to MR images, the denoising algorithm is applied, separatel y, to the two orthogonal sets of the complex MR image. The norm of the comb ined data are used to construct the image. With this method, signal-noise d ecoupling and Gaussian white noise assumptions used in the wavelet noise su ppression scheme, are better fulfilled. The performance of the method is te sted by carrying out a qualitative and quantitative comparison of a single- average image, complex-denoised image, multiple-average images, and a magni tude-denoised image, of a standard phantom. The comparison shows that the c omplex-denoising scheme improves the signal-to-noise and contrast-to-noise ratios more than the magnitude-denoising scheme, particularly in low SNR re gions. To demonstrate the method strength, it is applied to fMRI data of so matosensory rat stimulation. It is shown that the activation area in a cros s-correlation analysis is similar to 63% larger in the complex-denoised ver sus original data sets when equal threshold value is used. Application of t he method of Principal Component Analysis to the complex-denoised, magnitud e-denoised, and original data sets results in a similar but higher variance of the first few principal components obtained from the former data set as compared to those obtained from the later two sets. (C) 2000 Elsevier Scie nce Inc. All rights reserved.