A New Sequence for DTI with a Slab-Wise Diffusion Preparation, Reduced SAR and Increased Acquisition Speed.
Introduction
Diffusion Tensor Imaging (DTI, [1]) provides information regarding anatomical connections between cortical areas. It allows this to be done non-invasively in humans and at high fields with high spatial resolution. Here we introduce a new sequence with a slab-wise diffusion preparation. It consists of a 90 º pulse followed by a refocusing 180 º pulse and a -90 º to flip back the magnetization along the z axis. The diffusion gradients are placed around the 180 º pulse and subsequently provide encoding for all of the slices at one time, rather than once for each slice (Fig1, note that gradients are not drawn to scale). It is followed by N (here 15) slices of EPI readout. This sequence reduces SAR as well as acquisition time substantially.
Section snippets
Methods
One normal pineapple participated in this study. The experiments were performed at a 7T, 90 cm bore system consisting of a Magnex magnet and a Siemens console. Each acquisition consisted of the same Diffusion prepared 220 mm transverse slab going through the pineapple (see Fig1). It was subsequently read out by 15 transverse, interleaved GE EPI slices of 2 mm thickness each. A sixteen-channel geometrically adjustable volume coil was used [2]. (FOV = 19.2×19.2 cm2; matrix = 128×128; single shot
Results
Given the geometry of our pineapple phantom (Fig2) the reconstructed diffusion tensors look realistic (Fig3). From the 1050 tensors computed (Fig3), 358 had non-positive eigenvalues when computed with the LSM while none did with the constrained method. Just like the T2-weighted SPIF-T2 [4] this sequence is based on, this new sequence (SPIF-DIFF) reduces SAR significantly (more than 3-fold for 15 slices when compared to a standard multi-slice DTI sequence). Moreover it also more than doubles
Conclusions
This preliminary data set shows that DTI can be done with this new sequence. This method can now be expanded to more than 6 directions and to Q-ball type imaging [5]. It will be particularly suited for studies at Ultra High Field where low SAR and or high speed are a priority.