AUTHOR=He Junjie , Wang Lihui , Cao Ying , Wang Rongpin , Zhu Yuemin TITLE=Learn Less, Infer More: Learning in the Fourier Domain for Quantitative Susceptibility Mapping JOURNAL=Frontiers in Neuroscience VOLUME=Volume 16 - 2022 YEAR=2022 URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2022.837721 DOI=10.3389/fnins.2022.837721 ISSN=1662-453X ABSTRACT=Quantitative susceptibility mapping (QSM) aims to evaluate the distribution of magnetic susceptibility from magnetic resonance phase measurements by solving the ill-condition dipole inversion problem. Removing the artifacts and reserving the anisotropy of tissue susceptibility simultaneously is still a challenge in QSM. To deal with this issue, a novel k-QSM network is proposed to resolve dipole inversion issues in QSM reconstruction. The k-QSM converts the results obtained by truncated k-space division (TKD) into the Fourier domain as inputs, after several convolutional and residual blocks, the ill-posed signals of TKD are corrected by making the network output approach to the COSMOS-labeled QSM. For validating the superiority of k-QSM, comparisons with several state-of-the-art methods are performed in terms of QSM artifacts removing, anisotropy reserving, generalization ability and clinical applications. Compared to the best previous method, the k-QSM achieves a 9.41% lower NRMSE, 6.24% higher PSNR, 26.6% lower HFEN, and 0.65% higher SSIM. In addition, the orientation-dependent susceptibility variation obtained by k-QSM is the most significant, verifying that k-QSM has ability to reserve susceptibility anisotropy. When the trained models are tested on the dataset from different centers, our k-QSM shows a strong generalization ability with the highest PSNR. Moreover, by comparing the susceptibility maps between healthy controls and drug addicts with different methods, we found that k-QSM is more sensitive to the susceptibility change caused by some diseases. The proposed k-QSM method learns less, only to fix the ill-posed signals of TKD, but infers more, both COSMOS-like and anisotropy-reserving QSM results. Its generalization ability and great sensitivity to the susceptibility changes can make it a potential method for distinguishing some diseases.