Radio Frequency Coil Design for Ultra High Field Magnetic Resonance Imaging

多通道射频(RF)线圈阵列由多个谐振元件组成，是磁共振成像(MRI)采集中的关键组件。并行发射(PTX)阵列提供了单独的灵敏度分布，当与优化的梯度和RF波形配合使用时，可以加速激励K空间的遍历。这一原理可用于加速多维选择性激发，执行B1+或RF匀场以克服超高场(UHF)下的不均匀性，或降低比吸收率(SAR)。同样，并行接收(PRX)阵列利用表面线圈对MRI信号(B1-)的局部高信噪比(SNR)，并将其扩展到整个视场，同时执行空间编码，用于加速成像。在射频传输期间负责激发自旋的电磁场，以及从横向磁化后激励后接收信号的电磁场，随着主磁场强度的增加，从纯粹的反应性近场相互作用过渡到近场和远场的混合。正因为如此，与更传统的射频环路元件相比，电偶极子天线在UHF中的用途越来越大，环路和偶极阵列元件组合到单个RF线圈组件中预计将在单位B1+和SNR的SAR效率方面显示性能提升。然而，向使用电偶极子元件和/或组合不同元件的演变构成了一个巨大的技术挑战。由电偶极子产生的辐射方向图，连同它的几何形状，受到传统解耦方法的适用性的限制。这一点在人口稠密的偶极接收阵列中得到了最好的证明，在这些阵列中，未减弱的耦合和增强的噪声相关性会降低测量的SNR，而不考虑对样本的灵敏度增加。我们开发了一种通用解耦方法，允许使用梯形滤波器方法有效地对混合元件类型和可变耦合系数的复杂RF阵列进行解耦。使用这种新的方法，我们将模拟、设计和构建多通道发射阵列，这些阵列将环路和偶极子结合在一起，为人头成像提供尽可能多的发射效率。一旦硬件经过优化以通过射频垫片产生最佳的B1+均匀性，则将使用RF脉冲来消除残留的不均匀性。实时脉冲优化一直是一个挑战，因此目前使用折衷的方法。使用来自受试者和电磁模拟的B1+F和静态磁场图，我们将使用机器学习方法来帮助实时设计最优解决方案，该方法考虑到射频同质性和SAR的附加约束。这些方法结合在一起，将克服在临床和基础神经科学环境中为7T扫描仪采用多通道PTX的剩余障碍。

Multichannel radio frequency (RF) coil arrays, composed of multiple resonant elements, are a critical component in magnetic resonance imaging (MRI) acquisition. Parallel transmit (pTx) arrays provide individual sensitivity profiles that when used in concert with optimized gradient and RF waveforms can accelerate the traversal of excitation kspace. This principle can be used to accelerate multidimensional selective excitation, perform B1+ or RF shimming to overcome inhomogeneity at ultrahigh field (UHF), or reduce specific absorption rate (SAR). Similarly, parallel receive (pRx) arrays exploit the locally high signal-to-noise ratio (SNR) of surface coils to the MRI signal (B1-) and extend it across a full field of view while simultaneously performing spatial encoding, utilized in accelerated imaging.    The electromagnetic fields responsible for exciting spins during RF transmission, as well as receiving signal from the transverse magnetization post-excitation, transition from purely reactive nearfield interaction towards a mixture of both near and far fields as the main magnetic field strength increases. Due to this, electric dipole antennas are finding increasing utility at UHF when compared to more conventional RF loop elements and combinations of loop and dipole array elements into a single RF coil assembly are expected to show performance gains in SAR efficiency per unit B1+ and SNR.    The evolution towards using electric dipole elements and/or combining dissimilar elements present a technical challenge. The radiation patterns produced from the electric dipole, in conjunction with its geometry, are limited by the applicability of conventional decoupling methods. This is best demonstrated in densely populated dipole receive arrays where unmitigated coupling and enhanced noise correlations degrade measured SNR, regardless of an increased sensitivity to the sample. We have developed a generalized decoupling approach that allows complex RF arrays of mixed element types and variable coupling coefficients to be efficiently decoupled with a ladder filter approach. Using this approach, we will simulate, design and build multichannel transmit  arrays that combine loops and dipoles in a manner that captures as much of the transmit efficiency as possible for human head imaging. Once the hardware is optimized to generate the best B1+ homogeneity via RF shimming, RF pulses will be used to remove the residual inhomogeneity. Real-time pulse optimization has been a challenge, so compromised approaches are currently used. Using B1+ and static magnetic field maps from subjects and from electromagnetic simulations, we will use machine learning approaches to help design the optimal solutions in real-time that take into account RF homogeneity with the added constraints of SAR.    In combination, these approaches will surmount the remaining barriers to the adoption of multichannel pTx for 7 T scanners in clinical and basic neuroscience settings.