Noise Analysis in SMASH Imaging at Ultrahigh Field Magnetic Resonance Imaging

P.A. Wassenaar, P. Schmalbrock, and B.D. Clymer (USA)


Magnetic Resonance Imaging, Parallel Imaging, Image Restoration, Signal-to-Noise


Ultrahigh field magnetic resonance imaging allows the ac quisition of high resolution, high signal-to-noise ratio im ages. With the advantage of high resolution comes a cost of long acquisition times. Hence a desire to speed up acquisitions. This is complicated by the need to use transverse electromagnetic (TEM) resonators to deliver ra diofrequency (rf) fields at ultrahigh magnetic fields, which have coupled channels whose sensitivity regions overlap in the sample space. For the first time, a particular form of parallel imag ing is implemented on an ultrahigh field system: variable density autocalibrated simultaneous acquisition of spatial harmonics (VD-AUTO-SMASH). A reduced data set of k space is acquired to shorten scan time. These data are un dersampled in phase encoding directions, introducing alias ing, which appears as wrap-around artifacts in the corre sponding images. SMASH is an image restoration algo rithm that overcomes these aliasing effects, by using the difference in data gathered from different coil channels. VD-AUTO-SMASH uses some fully sampled data in a re gion near the origin of k-space to automatically calibrate the system to find the optimal reconstruction coefficients. It is shown that VD-AUTO-SMASH can be used at 8 Tesla (T) to reconstruct reduced data collected from a transverse electromagnetic (TEM) coil.

Important Links:

Go Back