Clipped DeepControl: Deep neural network two-dimensional pulse design with an amplitude constraint layer

• Published in: Artificial intelligence in medicine Vol. 135; p. 102460
• Main Authors: Vinding, Mads Sloth, Lund, Torben Ellegaard
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.01.2023
• Subjects: 2D RF; Clipping; DeepControl; Heart Rate; Magnetic Resonance Imaging - methods; MRI; Neural Networks, Computer; Optimal control; Phantoms, Imaging; Radio Waves; Clipping; 2D RF; MRI; DeepControl; Optimal control
• ISSN: 0933-3657; 1873-2860; 1873-2860
• DOI: 10.1016/j.artmed.2022.102460

Advanced radio-frequency pulse design used in magnetic resonance imaging has recently been demonstrated with deep learning of (convolutional) neural networks and reinforcement learning. For two-dimensionally selective radio-frequency pulses, the (convolutional) neural network pulse prediction time (a few milliseconds) was in comparison more than three orders of magnitude faster than the conventional optimal control computation. The network pulses were from the supervised training capable of compensating scan-subject dependent inhomogeneities of B0 and B1+ fields. Unfortunately, the network presented with a small percentage of pulse amplitude overshoots in the test subset, despite the optimal control pulses used in training were fully constrained. Here, we have extended the convolutional neural network with a custom-made clipping layer that completely eliminates the risk of pulse amplitude overshoots, while preserving the ability to compensate for the inhomogeneous field conditions. •DeepControl is an AI-powered source for MRI controls.•DeepControl computes advanced controls for MRI in milliseconds.•Clipped DeepControl prevents control amplitude overshoots.