Abstract #0785

DeepControl: AI-powered slice flip-angle homogenization by 2DRF pulses

Mads Sloth Vinding¹, Christoph Stefan Aigner², Jason Stockmann^3,4, Bastien Guérin^3,4, Sebastian Schmitter^2,5, and Torben Ellegaard Lund¹

¹Center of Functionally Integrative Neuroscience, Aarhus University, Aarhus, Denmark, ²Physikalisch-Technische Bundesanstalt (PTB), Braunschweig and Berlin, Germany, ³Harvard Medical School, Boston, MA, United States, ⁴Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, MA, United States, ⁵Center for Magnetic Resonance Research, University of Minnesota, Minneapolis, MN, United States

We test the DeepControl convolutional neural network for brain slice 2DRF 30^o excitations (single-channel) at 7 T with a uniform FA profile. While the DeepControl framework was originally designed for localized 2D spatial-selective excitations, we demonstrate robustness towards FA homogenization at 7 T. Our numerical study, a comparison to gradient-ascent-pulse-engineering optimal control pulses, shows good pulse performance and (near-)conformity to the optimal control training library, including the RF pulse constraints. The DeepControl pulse prediction time takes only ~9 ms, which is more than 1000 times faster than the optimal control.

This abstract and the presentation materials are available to members only; a login is required.

Join Here