You are here : Home > Research entities > Equipe Neurologique et Pharmacologique

Laboratory | Medical imaging

Clinical MRI methodological team

Parallel Transmission related activities

​​Team Leader: Alexandre Vignaud​

Published on 28 November 2017

NeuroSpin is equipped with a 8 channel transmission system. Up to now, our efforts were mainly focused on the none-selective RF excitation to homogenize the flip angle (small or large angle regimes) or the true rotation matrix on the human brain at 7T. Motivated by the success and simplicity of the Spoke method to settle uniform 2D excitation, our team has extended the concept to 3D though the kT-points method. The potential of the approach has been demonstrated first on density weighted imaging using spoiled gradient recalled echo. The RF pulse tailored for one specific subject played with 5 kT-points has proven to improve drastically flip angle spatial homogeneity at 7T on the human brain for a duration under the millisecond with a low energy deposition. The pulse parameterization has been combined with optimal control technique to tackle problem imposed by MP-RAGE inversion. RF inhomogeneity mitigation in this sequence is displayed on Figure 3 (CP defines standard transmission mode using Circular Polarisation) where white –gray matter (WM-GM) contrast is restored. 

Figure 3

In numerous applications, initial state is an unknown, which forbid to use small flip angle approximation classical approach to optimize kT-points. In this case, rotation angle and axis are specified. In this context, we have introduced GRadient Ascent Pulse Engineering (GRAPE) to synthetize desired rotation matrix for refocusing angles needed in 3D SPACE, Turbo Spin Echo at variable flip angles with a phase of rotation axis unleashed. Several decades of pulse are then designed using GPUs to implement 3D T2-weighted in vivo imaging. Results are presented on Figure 4.

Figure 4

Nevertheless, parallel transmission exploitation at UHF will reach its maximum potential only if the parameters which have a role on security are under control. Thus, our group has developed fast algorithms to design RF pulse, under SAR explicit constraints using Virtual Observation Points compression model, which allow to design 3D pulses in very competitive time (< 10 s for both small and large flip angles). In the context, the temperature would be a more relevant parameter than SAR. It has been demonstrated that the efficiency and the pulse duration could be improved if the temperature was taken into account instead of SAR. Unfortunately, it would require a good understanding of the SAR-temperature dependency in the biological tissues. The recent efforts from few groups to concentrate on this challenge  consists to measure temperature induced in vivo by a radiofrequency energy source with PRF (Proton Resonance Frequency shift) an thermometry MR method shown on Figure 5.

Figure 5