Accelerated MRI at 9.4 T with electronically modulated time-varying receive sensitivities
Parallel imaging (PI) is one of the most successful techniques for accelerating MRI acquisition. Based on the phased array concept, it exploits the spatial information provided by the different sensitivities of multiple local RF receive coils to complement the gradient-based Fourier encoding, thus speeding up the acquisition process. Since the sensitivity profiles of different coils are never completely spatially independent, PI reconstruction becomes increasingly ill-conditioned at higher acceleration factors, resulting in noise amplification and reconstruction artifacts. Therefore, efforts have been made in hardware design to improve coil arrays for optimized PI performance, in particular by increasing the number of elements.
In addition, the application of fast field modulations during image encoding has been proposed to improve accelerated imaging. In the present work, we investigate whether dynamic modulations of the receive RF coil sensitivities during image encoding can be used in a similar way to improve PI. To this end, we constructed a prototype 8-channel reconfigurable receive coil for human head imaging at ultra-high field strength of 9.4T, where sensitivity modulation is achieved by rapidly switching PIN diodes in the receive loops. These allow to change the surface current densities and thus to modulate the sensitivity patterns of the loops. With this setup, MR measurements were performed in both phantom and human subject. A reconstruction framework for PI with time-varying sensitivities was developed and applied to investigate how to best exploit this novel degree of freedom for image encoding. Lower reconstruction errors and noise amplification factors (~25% g-factor improvement for acceleration factor R=4) were demonstrated utilizing rapidly switched sensitivities compared to conventional reconstruction with static sensitivities.
In conclusion, this work introduced time-varying receive sensitivities enabled by fast-switching PIN diodes as a novel degree of freedom for spatial encoding and demonstrated the potential to improve PI at ultrahigh-field MRI.