Diffusion MRI Acquisition Without Shells

The patent extracts only the necessary diffusion direction samples needed for accurate microstructure parameter estimation, removing the redundant oversampling inherent in traditional shelled acquisitions. By using optimized sampling strategies and advanced reconstruction algorithms, the method obtains sufficient information from fewer diffusion directions, thereby reducing scan time while maintaining measurement precision for fiber orientation and microstructure characterization.

Instead of performing complete spherical shell sampling with 20-60 directions per scan parameter combination, the patent applies partial sampling by selecting a reduced set of diffusion directions that are optimally distributed. This partial action is sufficient when combined with advanced signal processing and reconstruction techniques, achieving the desired factorization of fiber orientation dispersion without the excessive time cost of full shell coverage.

The patent performs preliminary optimization of the diffusion direction sampling scheme before acquisition. By pre-calculating the optimal set of diffusion directions and their weights based on the desired rotational invariance properties, the method prepares a streamlined sampling plan that eliminates the need for excessive post-processing oversampling. This preliminary action enables efficient scanning across multiple relevant parameters while maintaining the ability to construct accurate rotational invariants.

The patent changes the sampling parameters from uniform shell-based distribution to optimized non-uniform distribution. By adjusting the diffusion directions, b-values, and timing parameters according to optimized trajectories and sampling patterns, the method achieves rotational invariance with fewer samples. This parameter optimization allows simultaneous sampling of multiple scan parameters (diffusion weightings, diffusion times, inversion and echo times) without being bottlenecked by direction oversampling.

The patent applies partial sampling by acquiring diffusion data from a reduced set of directions optimized for the specific microstructure estimation task. Rather than completing full spherical shells for every scan parameter combination, the method uses strategically selected diffusion directions that provide sufficient angular coverage when combined with advanced reconstruction algorithms. This partial action maintains adequate direction coverage while allocating more time per direction to improve signal-to-noise ratio.

The patent combines multiple sampling strategies and reconstruction approaches to create a composite acquisition and processing framework. By integrating optimized diffusion direction sampling with advanced signal processing, parallel imaging techniques, and iterative reconstruction methods, the method achieves comprehensive microstructure information from a composite of fewer, higher-quality measurements rather than numerous lower-quality measurements from complete shell coverage.