A Huygens' surface approach to rapid characterization of peripheral nerve stimulation. Issue 1 (24th August 2021)

• Record Type: Journal Article
• Authors: Davids, Mathias; Guerin, Bastien; Wald, Lawrence L.
• Is Part Of: Magnetic resonance in medicine. Volume 87:Issue 1(2022)
• Page Start: 377
• Page End: 393
• Publication Date: 2021-08-24
• Subjects: electromagnetic field simulation -- gradient coil design -- magneto‐stimulation thresholds -- MRI safety -- neurodynamic nerve model -- peripheral nerve stimulation; Nuclear magnetic resonance -- Periodicals; Electron paramagnetic resonance -- Periodicals; 616.07548
• DOI: 10.1002/mrm.28966
• Languages: English
• ISSNs: 0740-3194

Abstract

Abstract : Purpose: Peripheral nerve stimulation (PNS) modeling has a potential role in designing and operating MRI gradient coils but requires computationally demanding simulations of electromagnetic fields and neural responses. We demonstrate compression of an electromagnetic and neurodynamic model into a single versatile PNS matrix (P‐matrix) defined on an intermediary Huygens' surface to allow fast PNS characterization of arbitrary coil geometries and body positions. Methods: The Huygens' surface approach divides PNS prediction into an extensive pre‐computation phase of the electromagnetic and neurodynamic responses, which is independent of coil geometry and patient position, and a fast coil‐specific linear projection step connecting this information to a specific coil geometry. We validate the Huygens' approach by performing PNS characterizations for 21 body and head gradients and comparing them with full electromagnetic‐neurodynamic modeling. We demonstrate the value of Huygens' surface‐based PNS modeling by characterizing PNS‐optimized coil windings for a wide range of patient positions and poses in two body models. Results: The PNS prediction using the Huygens' P‐matrix takes less than a minute (instead of hours to days) without compromising numerical accuracy (error ≤ 0.1%) compared to the full simulation. Using this tool, we demonstrate that coils optimized for PNS at the brain landmark using a male model can also improve PNS for other imaging applications (cardiac, abdominal, pelvic, and knee imaging) in both male and female models. Conclusion: Representing PNS information on a Huygens' surface extended the approach's ability to assess PNS across body positions and models and test the robustness of PNS optimization in gradient design.