Isotropic diffusion encoding can be achieved via three separate orthogonal linear b-tensors (with successive dephasing and rephasing) in the context of triple diffusion encoding (see e.g., Mori and Van Zijl 1995, Wong et al. 1995 and Valette et al. 2012 for equivalent pulse sequences). However, faster spherical diffusion encoding can be achieved using optimized sequences such as the q-vector magic angle spinning (q-MAS) sequence introduced in Eriksson et al. 2013 or even more optimized gradient waveforms (see Sjölund et al. 2015 and Szczepankiewicz et al. 2020).
Of course, things are never that straightforward and optimized sequences come with an additional challenge: their frequency content. Indeed, optimized sequences rely on fast variations of the gradient amplitudes to achieve rapid diffusion encoding, which leads to higher frequencies in their encoding spectrum (see Lundell and Lasic 2020). This encoding spectrum indicates which frequencies of spin motions is probed by the acquisition. In particular, high encoding frequencies probe high-frequency (or short-time) spin motions, which can be affected by spatial restrictions to the diffusion process, as highlighted in Lundell et al. 2019. While such time-dependent effects seem to be rather small in the healthy brain, they could be crucial in diseased brain tissues. Outside of the brain, even healthy organs may already exhibit significant time-dependent effects.
This potential sensitivity to time-dependent diffusion could also be seen as a strength when combined with analysis methods that properly account for the link between the encoding spectrum and the probed spin-motion dynamics. Indeed, optimized spherical diffusion sequences tend to probe the highest frequencies and could be essential to future time-dependent studies, as recently proposed in the heart (Lundell et al. 2020).