Deep learning-based whole-brain B(1)(+)-mapping at 7T

Krueger, F.; Aigner, C.S.; Lutz, M.; Riemann, L.T.; Degenhardt, K.; Hadjikiriakos, K.; Zimmermann, F.F.; Hammernik, K.; Schulz-Menger, J.; Schaeffter, T.; Schmitter, S.

Deep learning-based whole-brain B(1)(+)-mapping at 7T

Tools

Preview	PDF (Original Article) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader 6MB
	Other (Supporting Information) 21MB

Item Type:	Article
Title:	Deep learning-based whole-brain B(1)(+)-mapping at 7T
Creators Name:	Krueger, F., Aigner, C.S., Lutz, M., Riemann, L.T., Degenhardt, K., Hadjikiriakos, K., Zimmermann, F.F., Hammernik, K., Schulz-Menger, J., Schaeffter, T. and Schmitter, S.
Abstract:	PURPOSE: This study investigates the feasibility of using complex-valued neural networks (NNs) to estimate quantitative transmit magnetic RF field (B(1)(+)) maps from multi-slice localizer scans with different slice orientations in the human head at 7T, aiming to accelerate subject-specific B(1)(+)-calibration using parallel transmission (pTx). METHODS: Datasets containing channel-wise B(1)(+)-maps and corresponding multi-slice localizers were acquired in axial, sagittal, and coronal orientation in 15 healthy subjects utilizing an eight-channel pTx transceiver head coil. Training included five-fold cross-validation for four network configurations: NN(tra)(cx) used transversal, NN(sag)(cx) sagittal, NN(cor)(cx) coronal data, and NN(all)(cx) was trained on all slice orientations. The resulting maps were compared to B(1)(+)-reference scans using different quality metrics. The proposed network was applied in-vivo at 7T in two unseen test subjects using dynamic kt-point pulses. RESULTS: Predicted B(1)(+)-maps demonstrated a high similarity with measured B(1)(+)-maps across multiple orientations. The estimation matched the reference with a mean relative error in the magnitude of (2.70 ± 2.86)% and mean absolute phase difference of (6.70 ± 1.99)° for transversal, (1.82 ± 0.69)% and (4.25 ± 1.62)° for sagittal (NN(sag)(cx)), as well as (1.33 ± 0.27)% and (2.66 ± 0.60)° for coronal slices (NN(cor)(cx)) considering brain tissue. NN(all)(cx) trained on all orientations enables a robust prediction of B(1)(+)-maps across different orientations. Achieving a homogenous excitation over the whole brain for an in-vivo application displayed the approach's feasibility. CONCLUSION: This study demonstrates the feasibility of utilizing complex-valued NNs to estimate multi-slice B(1)(+)-maps in different slice orientations from localizer scans in the human brain at 7T.
Keywords:	7 Tesla, B(1)(+), Mapping, Brain, Deep Learning, Parallel Transmission
Source:	Magnetic Resonance in Medicine
ISSN:	0740-3194
Publisher:	Wiley
Volume:	93
Number:	4
Page Range:	1700-1711
Date:	April 2025
Official Publication:	https://doi.org/10.1002/mrm.30359
PubMed:	View item in PubMed

Repository Staff Only: item control page

Download Statistics

Downloads

Downloads per month over past year