Echo‐planar imaging of the human head with 100 mT/m gradients and high‐order modeling of eddy current fields

Abstract

Purpose:To demonstrate the utility of a high-performance gradient insert for ultrafast MRI of the human head. Methods: EPI was used for the first time with a readout gradient amplitude of 100 mT/m, 1200 T/m/s slew rate, and nearly 1 MHz signal bandwidth for human head scanning. To avoid artefacts due to eddy currents, the magnetic field was dynamically monitored with NMR probes at multiple points, modeled by solid harmonics up to fifth order, and included in the image reconstruction. An approximation of a negligible intra-echo effect of the eddy currents was made to accelerate the high-order reconstruction. The field monitoring-based approach was compared with a recently proposed phase error estimation from separate reconstructions of even and odd echoes. Results: Images obtained with the gradient insert have significantly lower distortions than it is the case with the whole body 30 mT/m, 200 T/m/s gradients of the same system. However, eddy currents of high spatial order must be properly characterized and corrected for in order to avoid a persistent 2D Nyquist ghost. Multi-position monitoring proves to be a robust method to measure the eddy currents and allows higher undersampling rates than the image-based approach. The proposed approximation of the eddy currents effect allows a significant acceleration of the high-order reconstruction by a separate processing of each spatial dimension. Conclusion: Strong gradients with adequate switching rates are highly beneficial for the quality of EPI provided that robust measures are taken to include the contribution of eddy currents to the image encoding. K E Y W O R D Seddy currents, EPI, gradients, higher-order encoding, nyquist ghost

Minimizing the echo time in diffusion imaging using spiral readouts and a head gradient system

“…The magnitude of the phase variation during the readout was around 10 radians within the imaging volume. Related eddy currents with time constants of a few milliseconds or less were previously shown to result in ghosting artifacts in EPI 27 . In the case spiral imaging, the related artifacts appear as blurring, which may seem less critical at first sight (Figure 5A).…”

“…From the combined data, a fifth‐order spherical harmonic model was fitted. This has previously been reported to sufficiently describe eddy currents generated by this gradient coil to allow for EPI imaging 27 . The choice of spherical harmonics as basic functions to describe magnetic field dynamics is motivated by magnetostatics.…”

“…This has previously been reported to sufficiently describe eddy currents generated by this gradient coil to allow for EPI imaging. 27 The choice of spherical harmonics as basic functions to describe magnetic field dynamics is motivated by magnetostatics. However, the effect of the concomitant fields on the magnetic field magnitude cannot be fully described by this basis, and therefore, may be modeled separately.…”

15 Years MR-encephalography

“…To facilitate the investigation of fast BOLD fluctuations, it was then natural to combine MREG with simultaneous EEG, whose high temporal resolution would provide useful baseline recordings of cerebral activity. While simultaneous EEG-fMRI had been previously almost exclusively used with EPI, it was quickly established that fast acquisitions and spiral trajectories did not cause additional safety issues [40,73], nor prevent an effective removal of MR gradient artifacts from the EEG [74,75]. As a starting point, we focused on the default-mode network (DMN), which is associated with the brain's baseline state, giving it a central role as a hub of global brain activity [76].…”

“…Most recently other groups have been working on ultrafast spiral imaging. A repetition time of 200 ms at 2.8 mm isotropic resolution has been reported using a T-Hex singleshot spiral trajectory [39] and 3.5 ms/slice could be achieved with 2D-spiral imaging using a dedicated gradient coil [40].…”