Deep attentive spatio-temporal feature learning for automatic resting-state fMRI denoising

Heo, Keunsoo; Shin, Dong-Hee; Hung, Sheng-Che; Lin, Weili; Zhang, Han; Shen, Dinggang; Kam, Tae-Eui; Consortium, For Unc/Umn Baby Connectome Project

doi:10.1016/j.neuroimage.2022.119127

Cited by 7 publications

(5 citation statements)

References 90 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Future work aims to improve the image quality of reconstructed images in low–SNR scenarios. One potential solution is to construct a noise estimation network to fully characterize the noise in the fMRI signal [ 51 , 52 ]. By removing the noise properly, it can enhance the model’s performance from a data perspective.…”

Section: Discussionmentioning

confidence: 99%

Natural Image Reconstruction from fMRI Based on Node–Edge Interaction and Multi–Scale Constraint

Kuang,

Zhan,

Gao

2024

Brain Sciences

View full text Add to dashboard Cite

Reconstructing natural stimulus images using functional magnetic resonance imaging (fMRI) is one of the most challenging problems in brain decoding and is also the crucial component of a brain–computer interface. Previous methods cannot fully exploit the information about interactions among brain regions. In this paper, we propose a natural image reconstruction method based on node–edge interaction and a multi-scale constraint. Inspired by the extensive information interactions in the brain, a novel graph neural network block with node–edge interaction (NEI–GNN block) is presented, which can adequately model the information exchange between brain areas via alternatively updating the nodes and edges. Additionally, to enhance the quality of reconstructed images in terms of both global structure and local detail, we employ a multi-stage reconstruction network that restricts the reconstructed images in a coarse-to-fine manner across multiple scales. Qualitative experiments on the generic object decoding (GOD) dataset demonstrate that the reconstructed images contain accurate structural information and rich texture details. Furthermore, the proposed method surpasses the existing state-of-the-art methods in terms of accuracy in the commonly used n-way evaluation. Our approach achieves 82.00%, 59.40%, 45.20% in n-way mean squared error (MSE) evaluation and 83.50%, 61.80%, 46.00% in n-way structural similarity index measure (SSIM) evaluation, respectively. Our experiments reveal the importance of information interaction among brain areas and also demonstrate the potential for developing visual-decoding brain–computer interfaces.

show abstract

Section: Discussionmentioning

confidence: 99%

Natural Image Reconstruction from fMRI Based on Node–Edge Interaction and Multi–Scale Constraint

Kuang,

Zhan,

Gao

2024

Brain Sciences

View full text Add to dashboard Cite

show abstract

“…Even for a single time point, deep learning can capture latent information through up-sampling methods [93]. Given the low signal-to-noise rate and repetition of fMRI data [97], the deep-learning denoising network [98] has the potential to be integrated into frameworks to obtain reliable and robust parcellations. Furthermore, the excellent processing of temporal or contextual information of deep learning will also provide new perspectives on the dynamics of brain parcellation.…”

Section: Hierarchy Dynamic and Multimodalitymentioning

confidence: 99%

Application of deep learning in fMRI-based human brain parcellation: a review

Li,

Chen,

Ling

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

Functional magnetic resonance imaging (fMRI)-based human brain parcellation reveals brain fundamental organizational principles noninvasively, providing prior guidance for functional analysis and physiological measurement of the brain. Recently, the profound success of deep learning in natural and medical images has attracted an increasing number of deep learning-based investigations for brain parcellation which have accomplished encouraging achievements. This review aims to provide researchers with a comprehensive overview of deep learning-based fMRI brain parcellation and promote the development of this emerging frontier. To begin, we present the history of brain parcellation, emphasizing the current advantages of deep learning in this field. Following a summary of the input modalities, traditional methods, and evaluation criteria for fMRI-based brain parcellation, we comprehensively review the current deep-learning approaches and categorize them based on network architectures. The datasets, models, performance, advantages and limitations are described in detail. Finally, we discuss the current challenges and future directions in methodology, neuroscience, and application, to inspire future investigation in deep learning-based brain parcellation.

show abstract

“…Lin et al used a deep image prior (DIP) to simultaneously denoise all diffusion-weighted images. The method demonstrated superior performance when compared with the local principal component analysis method using both simulated and in vivo datasets [ 129 ]. Kawamura et al evaluated the application of CNN-based denoising for multi-shot EPI DWI and compared the deep denoiser with other methods including block-matching and 3D filtering [ 130 ].…”

Section: Noise Reduction In Mrimentioning

confidence: 99%

Deep Learning for Image Enhancement and Correction in Magnetic Resonance Imaging—State-of-the-Art and Challenges

et al. 2022

View full text Add to dashboard Cite

Magnetic resonance imaging (MRI) provides excellent soft-tissue contrast for clinical diagnoses and research which underpin many recent breakthroughs in medicine and biology. The post-processing of reconstructed MR images is often automated for incorporation into MRI scanners by the manufacturers and increasingly plays a critical role in the final image quality for clinical reporting and interpretation. For image enhancement and correction, the post-processing steps include noise reduction, image artefact correction, and image resolution improvements. With the recent success of deep learning in many research fields, there is great potential to apply deep learning for MR image enhancement, and recent publications have demonstrated promising results. Motivated by the rapidly growing literature in this area, in this review paper, we provide a comprehensive overview of deep learning-based methods for post-processing MR images to enhance image quality and correct image artefacts. We aim to provide researchers in MRI or other research fields, including computer vision and image processing, a literature survey of deep learning approaches for MR image enhancement. We discuss the current limitations of the application of artificial intelligence in MRI and highlight possible directions for future developments. In the era of deep learning, we highlight the importance of a critical appraisal of the explanatory information provided and the generalizability of deep learning algorithms in medical imaging.

show abstract

Deep attentive spatio-temporal feature learning for automatic resting-state fMRI denoising

Cited by 7 publications

References 90 publications

Natural Image Reconstruction from fMRI Based on Node–Edge Interaction and Multi–Scale Constraint

Natural Image Reconstruction from fMRI Based on Node–Edge Interaction and Multi–Scale Constraint

Application of deep learning in fMRI-based human brain parcellation: a review

Deep Learning for Image Enhancement and Correction in Magnetic Resonance Imaging—State-of-the-Art and Challenges

Contact Info

Product

Resources

About