Principal Investigator: Prof Ernesta Meintjes

Through our research we have provided training and capacity building in this highly specialized area, both in terms of training people in advanced image analysis for clinical research and in MRI technology.

In collaboration with Andre van der Kouwe from Massachusetts General Hospital (MGH), we have developed a suite of sequences optimized for paediatric neuroimaging that include real-time motion, and for spectroscopy B0, correction using volumetric navigators. These have included sequences for both single (Hess et al., 2011) and multivoxel (Hess et al., 2012) MRS, structural MRI (Tisdall et al., 2012), and DTI (Alhamud et al., 2012).

During cardiac magnetic resonance imaging, patients often need to hold their breath for extended periods. Traditionally, navigator methods only use data acquired when the diaphragm position falls within a pre-specified window, resulting in low respiratory efficiency (40%) and long scan times. We developed a novel method to instead measure with a navigator the position of the diaphragm during non-imaging segments, and a control system to predict the diaphragm position during imaging segments (Burger et al., 2010). A subject-specific nonlinear elliptical affine model that incorporates the effect of hysteresis in motion correction and can be constructed from 25s of pre-scan data, has been developed to predict the heart position based on the diaphragm position outputs from the navigator (Burger et al., 2013). By combining these two methods, the imaging slice position can be updated in real time during free-breathing CMR resulting in 100% respiratory efficiency.