MMR-Mamba: Multi-Modal MRI Reconstruction with Mamba and Spatial-Frequency Information Fusion

TL;DR

MMR-Mamba addresses the challenge of reconstructing a target MRI modality from under-sampled data by leveraging a fully-sampled reference modality. It combines a Target modality-guided Cross Mamba (TCM) for efficient spatial fusion with a Selective Frequency Fusion (SFF) to recover global frequency information, and further enhances cross-domain synergy via Adaptive Spatial-Frequency Fusion (ASFF). The approach demonstrates superior reconstruction quality on BraTS and fastMRI knee datasets, with ablations confirming the contribution of each module. This framework offers a practical, computation-efficient pathway for high-quality multi-modal MRI reconstruction, enabling faster acquisitions without sacrificing structural fidelity. Overall, MMR-Mamba advances multi-modal MRI reconstruction by uniting linear-complexity long-range modeling with principled spatial-frequency information fusion, delivering robust improvements across challenging acceleration factors.

Abstract

Multi-modal MRI offers valuable complementary information for diagnosis and treatment; however, its utility is limited by prolonged scanning times. To accelerate the acquisition process, a practical approach is to reconstruct images of the target modality, which requires longer scanning times, from under-sampled k-space data using the fully-sampled reference modality with shorter scanning times as guidance. The primary challenge of this task is comprehensively and efficiently integrating complementary information from different modalities to achieve high-quality reconstruction. Existing methods struggle with this: 1) convolution-based models fail to capture long-range dependencies; 2) transformer-based models, while excelling in global feature modeling, struggle with quadratic computational complexity. To address this, we propose MMR-Mamba, a novel framework that thoroughly and efficiently integrates multi-modal features for MRI reconstruction, leveraging Mamba's capability to capture long-range dependencies with linear computational complexity while exploiting global properties of the Fourier domain. Specifically, we first design a Target modality-guided Cross Mamba (TCM) module in the spatial domain, which maximally restores the target modality information by selectively incorporating relevant information from the reference modality. Then, we introduce a Selective Frequency Fusion (SFF) module to efficiently integrate global information in the Fourier domain and recover high-frequency signals for the reconstruction of structural details. Furthermore, we devise an Adaptive Spatial-Frequency Fusion (ASFF) module, which mutually enhances the spatial and frequency domains by supplementing less informative channels from one domain with corresponding channels from the other.

Figures (5)

• Figure 1: Overview of the proposed MMR-Mamba framework (left). It contains Mamba blocks for feature extraction, TCM for spatial domain fusion, SFF for frequency domain fusion, and ASFF for spatial-frequency information integration. Structure of Mamba block and TCM (right).
• Figure 2: Illustration of Selective Frequency Fusion (SFF) module.
• Figure 3: Illustration of the Adaptive Spatial-Frequency Fusion module.
• Figure 4: Qualitative evaluation of reconstruction results from different methods on BraTS dataset and fastMRI knee dataset under 4$\times$ and 8$\times$ acceleration. For every group, the first row shows the reconstructed images and the second row displays the error map between the results and the ground truth. More color in the error map indicates worse reconstruction results.
• Figure 5: Visualization of the results from ablation study of our proposed modules on BraTS dataset under 8$\times$ acceleration.