Numerous clinical and research applications for quantitative mapping of the effective transverse relaxation time T*(2) have been described. Subject motion can severely deteriorate the quality and accuracy of results. A correction method for T*(2) maps acquired with multi-slice multiple gradient echo FLASH imaging is presented, based on acquisition repetition with reduced spatial resolution (and consequently reduced acquisition time) and weighted averaging of both data sets, choosing weighting factors individually for each k-space line to reduce the influence of motion. In detail, the procedure is based on the fact that motion artifacts reduce the correlation between acquired and exponentially fitted data. A target data set is constructed in image space, choosing the data yielding best correlation from the two acquired data sets. The k-space representation of the target is subsequently approximated as linear combination of original raw data, yielding the required weighting factors. As this method only requires a single acquisition repetition with reduced spatial resolution, it can be employed on any clinical system offering a suitable sequence with export of modulus and phase images. Experimental results show that the method works well for sparse motion, but fails for strong motion affecting the same k-space lines in both acquisitions.
Copyright © 2011 Wiley-Liss, Inc.