Frequency and phase correction of J‐difference edited MR spectra using deep learning

Abstract

To investigate whether a deep learning-based (DL) approach can be used for frequency-and-phase correction (FPC) of MEGA-edited MRS data. Methods: Two neural networks (1 for frequency, 1 for phase) consisting of fully connected layers were trained and validated using simulated MEGA-edited MRS data. This DL-FPC was subsequently tested and compared to a conventional approach (spectral registration [SR]) and to a model-based SR implementation (mSR) using in vivo MEGA-edited MRS datasets. Additional artificial offsets were added to these datasets to further investigate performance. Results: The validation showed that DL-based FPC was capable of correcting within 0.03 Hz of frequency and 0.4°of phase offset for unseen simulated data. DL-based FPC performed similarly to SR for the unmanipulated in vivo test datasets. When additional offsets were added to these datasets, the networks still performed well. However, although SR accurately corrected for smaller offsets, it often failed for larger offsets. The mSR algorithm performed well for larger offsets, which was because the model was generated from the in vivo datasets. In addition, the computation times were much shorter using DL-based FPC or mSR compared to SR for heavily distorted spectra. Conclusion: These results represent a proof of principle for the use of DL for preprocessing MRS data.

MR spectroscopy frequency and phase correction using convolutional neural networks

“…In vivo data were retrieved from the publicly available Big GABA repository 8 . Thirty‐three MEGA‐edited datasets were collected in total.…”

“…Since there is no ground truth of frequency and phase offsets for the in vivo dataset, in this work, the MEGA-PRESS training, validation, and test transients were simulated using the FID-A toolbox (version 1.2), with the same parameters as described in the previous work. 8 The training set for the CNN model was allocated 36,000 OFF+ON spectra, the validation set was allocated 4,000, and 1,000 for the test set. Furthermore, we created additional spectra with lower SNRs (10, 5, and 2.5) by adding random Gaussian noise to the published simulated dataset, respectively.…”

“…9 As another comparison for our model, a benchmark neural network using MLP containing three FC layers (1024, 512, 1 node(s)) was tested. 8 In this network, each hidden FC layer was followed by a ReLU activation function, and a linear activation function followed the output layer.…”

Magnetic Resonance Spectroscopy Frequency and Phase Correction Using Convolutional Neural Networks

“…In vivo data was retrieved from the publicly available Big GABA repository [10]. Thirty-three MEGA-edited datasets were collected in total.…”

ORYX-MRSI: A Fully-Automated Open-Source Software for Three-Dimensional Proton Magnetic Resonance Spectroscopic Imaging Data Analysis

“…(Kantarci et al, 2000;Nelson, 2003Nelson, , 2011 Despite the vast amount of information provided by 1 H-MRSI, it is still not widely employed in the clinical settings. As a result, there has been a major effort for improving the clinical utility of 1 H-MRSI with recent developments in data acquisition, processing, and quantitative analysis aspects (Bartnik-Olson et al, 2021;Chiew et al, 2018;Hingerl et al, 2020;Landheer et al, 2020;Near et al, 2021;Oeltzschner et al, 2019;Oz et al, 2020;Povazan et al, 2020;Tapper et al, 2021;Wilson et al, 2019). As part of these extensive efforts, open-source command-line scripts or software with user-friendly graphical user interfaces (GUIs) have been released in the past few year (Clarke et al, 2021;Crane et al, 2013;Edden et al, 2014;Maudsley et al, 2006Maudsley et al, , 2009Naressi et al, 2001;Oeltzschner et al, 2020;Poullet et al, 2007;Provencher, 1993;Reynolds et al, 2006;Simpson et al, 2017;Soher et al, 2011;Wilson et al, 2011;Wilson, 2021).…”