Shape Optimization of an Electric Dipole Array for 7 Tesla Neuroimaging

• Published in: IEEE transactions on medical imaging Vol. 38; no. 9; pp. 2177 - 2187
• Main Authors: Connell, Ian R. O., Menon, Ravi S.
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.09.2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Arrays; Brain - diagnostic imaging; Brain slice preparation; Conductors; coupling matrix synthesis; Couplings; Design optimization; dipoles; Electric dipoles; Equipment Design; Head; Head - diagnostic imaging; Humans; Image Processing, Computer-Assisted; Impedance; Magnetic resonance imaging; Magnetic Resonance Imaging - instrumentation; Magnetic Resonance Imaging - methods; Medical imaging; MRI RF coils; Neuroimaging; Neuroimaging - methods; Neurology; Noise levels; Optimization; Orthogonality; Phantoms, Imaging; Radio frequency; Radio-frequency arrays; RF arrays; RF engineering; Shape; Shape optimization; transceiver arrays; ultra-high-field MRI
• ISSN: 0278-0062; 1558-254X; 1558-254X
• DOI: 10.1109/TMI.2019.2906507

Radio-frequency (RF) arrays constructed using electric dipoles have potential benefits for transmit and receive applications using the ultra-high field (UHF) MRI. This paper examines some of the implementation barriers regarding dipole RF arrays for human head imaging at 7 T. The dipole array was constructed with conformal, meandered dipoles with dimensions selected utilizing an evolutionary-based optimization routine to shape-optimize the dipole structure. Coupling matrix synthesis (CMS) was utilized to decouple the dipole array. Mean and worst-case transmission between nearest-neighbour dipoles was −17.2 and −15.5 dB, respectively (±2.4 dB). Transmit efficiencies of 24.6 nT/V for the entire brain and 26.0 nT/V across the axial slice were observed. The total and peak 10-g SAR, normalized to 1 Watt accepted input power per channel, was 0.163 and 0.601 W/kg, respectively. Maximum and mean noise correlations were −17 dB and −32 dB, respectively. The use of both CMS and a novel shape optimization routine to design a dipole array translated into sufficient transmit uniformity with a simultaneous reduction in 10-g SAR in comparison to a non-optimized dipole array of the same geometry. As a receiver, the dipole array maintained high orthogonality between elements, resulting in strong parallel imaging performance.