MRI Drive Sequence Generation Using Neural Network Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current magnetic resonance imaging (MRI) techniques face challenges in achieving target excitation states, particularly at ultra-high field strengths, due to spatial inhomogeneity of the radiofrequency excitation field and lengthy optimization processes, which result in delayed examination readiness and limited correction of field inhomogeneities.
Innovation Solution
A method that precalculates drive sequences for specific reference points, using interpolation and extrapolation with a trained neural network to quickly generate drive sequences suitable for target excitation states, incorporating complex radiofrequency and gradient pulses to ensure short output duration and high spatial selectivity, even at ultra-high field strengths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If optimization methods are employed to determine optimization parameters for dynamic and multi-pulse pTx drive sequences, then target excitation states can be achieved, but the process becomes extremely complex and computationally intensive, resulting in very long delays
Solution Approach 1:
The patent precalculates base sequences for a discrete set of reference points covering the parameter space before the actual examination. These precomputed base sequences are stored and later used with interpolation/extrapolation to quickly generate drive sequences for specific measurement procedures, eliminating the need for time-consuming optimization during clinical examinations.
Solution Approach 2:
The patent segments the continuous parameter space into discrete reference points and organizes base sequences into clusters. By dividing the optimization problem into precomputation at reference points and runtime interpolation, the system achieves both accuracy and speed.
2Adaptability or versatility
If precalculation of base sequences is performed for all conceivable target excitation parameters, then complete coverage is achieved, but the precalculation time extends to several months or years
Solution Approach 1:
The patent divides the continuous parameter space into discrete reference points and organizes them into clusters. Only base sequences at these discrete reference points are precalculated, not the entire continuous parameter space. This segmentation reduces precalculation time from months/years to a manageable duration while maintaining versatility through interpolation and extrapolation for intermediate values.
Solution Approach 2:
The patent creates universal base sequences at reference points that can serve multiple purposes through interpolation and extrapolation. These precomputed sequences are not limited to their specific reference parameters but can generate drive sequences for any parameter within the covered range, achieving multi-functionality.
3Measurement precision
If higher magnetic field strengths of 7 Tesla or more are employed, then higher image quality is achieved, but spatial inhomogeneity of the radiofrequency excitation field increases due to short Larmor wavelength
Solution Approach 1:
The patent changes the temporal parameters of radiofrequency pulses by applying time-varying modulations to compensate for spatial inhomogeneities. Through dynamic parallel transmission with time-varying B1 fields and coordinated gradient fields, the system achieves homogeneous flip-angle distributions despite the inherent spatial variations at ultra-high field strengths.
Solution Approach 2:
The patent employs dynamic parallel transmission where radiofrequency pulse shapes and gradient fields vary over time. This dynamic approach allows continuous adjustment of the excitation field to maintain homogeneity across the target region, contrasting with static transmission methods that cannot adapt to spatial variations.
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 rapid generation of drive sequences that achieve high-quality target excitation states with improved spatial selectivity and reduced precalculation times, making MRI examinations more clinically feasible by minimizing delays and optimizing field homogeneity.
Implementation Method 1
A radiofrequency coil arrangement has a plurality of radiofrequency coils that can be driven via respective transmit channels with radiofrequency pulses
Implementation Method 2
In magnetic resonance imaging (e.g., also with a medical use), it has been proposed to employ higher magnetic field strengths of the main magnetic field (B0 field)
Implementation Method 3
A gradient coil arrangement has a plurality of gradient coils that can be driven via respective gradient channels with gradient pulses
Data Source
AI summary
A computer-implemented method for providing a drive sequence for a magnetic resonance device is provided. The drive sequence includes radiofrequency pulses to be output via transmit channels of a radiofrequency coil arrangement. The method includes, before the examination including the measurement procedure, precalculating a set of base sequences for mutually spaced reference points of at least one requirement parameter that describes the target excitation state. The reference points cover a parameter interval for use. The base sequences are provided together with the associated reference points at the magnetic resonance device. A measurement procedure value of the requirement parameter is provided at the magnetic resonance device. A drive sequence is ascertained for use for the measurement procedure. In the event that the measurement procedure value for the requirement parameter differs from the reference points, the drive sequence is ascertained from the base sequences by interpolation and/or extrapolation using a derivation algorithm.