MRSI Reconstruction via Sparse (k,t)-Space Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance spectroscopic imaging (MRSI) faces challenges with long data acquisition times, poor spatial resolution, and low signal-to-noise ratio (SNR), limiting its clinical and research applications.
Innovation Solution
A method and apparatus that utilize sparse sampling in (k,t)-space with variable density and subspace models to acquire and reconstruct high-resolution MRSI data, enabling high-speed data acquisition while maintaining good SNR, by organizing data into sets with high SNR and temporal resolution, and extended k-space coverage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional MRSI data acquisition methods are used, then spectral information is obtained, but data acquisition time is long
Solution Approach 1:
The patent applies partial sampling in k-space and t-space, acquiring only a subset of the full data matrix (e.g., 10-50% of k-space lines and 10-50% of time points). This partial action reduces acquisition time while subspace reconstruction algorithms recover the missing information, achieving speedup without complete data loss
Solution Approach 2:
The patent performs preliminary organization of acquired data into training sets and imaging sets before reconstruction. The training set is used to establish subspace models in advance, which then guide the reconstruction of the full spatiospectral function, enabling faster processing of the imaging data
2Measurement precision
If MRSI is performed with adequate spatial resolution, then metabolic information is captured, but signal-to-noise ratio is low
Solution Approach 1:
The patent transitions from traditional 2D spatial encoding to 3D spatiospectral encoding by adding the spectral dimension (frequency/chemical shift) as a third encoding dimension. This creates a (kx, ky, ω) space that captures spatial and metabolic information simultaneously, improving measurement precision without proportionally increasing noise
Solution Approach 2:
The patent changes the sampling parameters in k-space and t-space, using variable density sampling where central k-space regions are sampled more densely than peripheral regions. It also optimizes the number of time points and k-space lines to balance resolution and SNR based on the specific application requirements
3Productivity
If fast data acquisition is implemented using echo-planar spectroscopic imaging, then data acquisition time is reduced, but signal-to-noise ratio deteriorates
Solution Approach 1:
The patent uses partial sampling of k-space and t-space with subspace reconstruction to achieve acceleration factors of 2-10x compared to traditional methods. By sampling only essential data points and reconstructing the remainder through low-rank modeling, it maintains SNR while dramatically improving acquisition speed
Solution Approach 2:
The patent employs iterative reconstruction algorithms that use feedback from the acquired data to refine the spatiospectral function estimate. The reconstruction process continuously adjusts the model to match the measured data while enforcing subspace constraints, improving SNR through multiple refinement cycles
4Productivity
If parallel imaging with phased array coils is used, then data acquisition is accelerated, but system complexity increases
Solution Approach 1:
The patent accelerates data acquisition through mathematical sampling strategies (partial k-space and t-space sampling with subspace reconstruction) rather than adding physical hardware. This achieves speedup without increasing device complexity, using software-based reconstruction instead of additional coils or hardware components
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach allows for high-speed, high-resolution MRSI data acquisition, improving spatial resolution and SNR, thereby enhancing the effectiveness of MRSI in providing biochemical information.
Implementation Method 1
When a substance such as human tissue is subjected to a uniform magnetic field (such as a polarizing field B0), the individual magnetic moments of the spins in the tissue attempt to align with this polarizing field, but precess about it in random order at their characteristic Larmor frequency.
Implementation Method 2
If the substance, or tissue, is subject to a magnetic field (excitation field B1) in the x-y plane and which is near the Larmor frequency, the net aligned moment, or 'longitudinal magnetization', Mz, may be rotated or 'tipped' into the x-y plane to produce a net transverse magnetic moment Mt.
Implementation Method 3
A signal is emitted by the excited spins after the excitation signal B1 is terminated and this signal may be received and processed to form an image.
Implementation Method 4
When utilizing these signals to produce images, magnetic field gradients (Gx, Gy, and Gz) are employed. Typically, the region to be imaged is scanned by a sequence of measurement cycles in which these gradients vary according to the particular localization method used.
Data Source
AI summary
Various embodiments accelerate high-resolution magnetic resonance spectroscopic imaging (MRSI). Various embodiments are built on a low-dimensional subspace model exploiting the partial separability of high-dimensional MRSI signals. For two and three dimensional MRSI with one spectral dimension, various embodiments sparsely sample the corresponding (k,t)-space in two complementary data sets, one with dense temporal sampling and high signal-to-noise ratio but limited k-space coverage and the other with sparse temporal sampling but extended k-space coverage. The reconstruction is then done by estimating a set of temporal/spectral basis functions and the corresponding spatial coefficients from these two data sets. The imaging technique of various embodiments can be used for high-resolution MRSI of different nuclei.