Multiphoton parallel Transmit for MRI

7. Project Summary/Abstract: Multiphoton parallel transmit for MRI Magnetic Resonance Imaging (MRI) has been the undisputed standard of care for detecting and diagnosing neurological disorders and injuries, and 7 Tesla MRI has repeatedly shown its potential for increasing contrast and resolution. However, translation to the clinic has been hampered by the flip angle inhomogeneity problem in high-field MRI. This results from wave interference effects that cause constructive interferences of the B1+ fields at the head's center and destructive interferences at the periphery. Perhaps the most promising mitigation approach has been B1 shimming or full parallel transmit (pTx). However, a full pTx system adds considerable costs (extra RF amplifier, cabling, and power monitoring equipment) as well as complicated the estimation of local SAR, since maps of the electric field vector (E) for each channel must be known and vectorially added to determine the total local electric field and, thus, SAR. The operator's confidence in the local SAR predictions will depend on how well patient anatomy and positioning match the population of numerical body models used to predict the electric fields. Additionally, the RF waveforms must be continuously monitored during the scan to ensure they follow the prescribed waveforms. Lack of confidence in the local SAR estimation procedure necessitates conservative power restrictions to ensure subject safety. We address the hardware and SAR complications of pTx by introducing an excitation scheme that utilizes a conventional birdcage RF coil, but achieves the benefits of conventional pTx by supplementing the excitation with multiple low-frequency (kHz) transmitters placed around the head. We employ the multiphoton MR phenomenon, which consists of an off-resonant excitation from a conventional birdcage coil together with low- frequency, sinusoidal, localized, and spatially nonlinear variations in the z-directed magnetic field generated by a multichannel shim array to shape the magnetization distribution. We formulate the multiphoton pulse optimization problem in the small-tip regime to find multichannel shim array current amplitudes and phases that substantially improve the transverse magnetization homogeneity in a brain region-of-interest (ROI). Using the multiphoton phenomenon as a “correction” pulse following a conventional on-resonance birdcage excitation, we generate highly uniform flip angle distributions that reduce the transverse magnetization magnitude NRMSE over the brain four-fold compared to conventional birdcage excitation at 7 T in pulses that respect gradient slew limits and have a total length of 1 ms. We will convert our shim array apparatus to provide the needed z-directed oscillating fields and validate this approach for 7 Tesla brain imaging phantom and human studies.