Retrospective Head Motion Estimation in Structural Brain MRI with 3D CNNs (original) (raw)
Related papers
2018
Head motion during MRI acquisition presents significant problems for subsequent neuroimaging analyses. In this work, we propose to use convolutional neural networks (CNNs) to correct motion-corrupted images as well as investigate a possible improvement by augmenting L1 loss with adversarial loss. For training, in order to gain access to a ground-truth, we first selected a large number of motion-free images from the ABIDE dataset. We then added simulated motion artifacts on these images to produce motion corrupted data and a 3D regression CNN was trained to predict the motion-free volume as the output. We tested the CNN on unseen simulated data as well as real motion affected data. Quantitative evaluation was carried out using metrics such as Structural Similarity (SSIM) index, Correlation Coefficient (CC), and Tissue Contrast T-score (TCT). It was found that Gaussian smoothing as a conventional method did not significantly differ in SSIM, CC and RMSE from the uncorrected data. On th...
Signal-to-noise ratio estimates predict head motion presence in T1-weighted MRI
MRIQC (Esteban et al. 2017) is a tool to help researchers perform quality control (QC) on their structural and functional MRI data. Not only does MRIQC generate visual reports for reliable, manual assessment but it also automatically extracts a set of image quality metrics (IQMs). However, these IQMs are hard to interpret, and many related questions remain open, such as which IQMs are more important. In this project, which emerged as a BrainHack Geneva 2022 initiative, we show that head motion during the acquisition of whole-brain T1-weighted (T1w) MRI of healthy volunteers can be predicted based on the IQMs using supervised machine learning. To do so, we employ the open MR-ART (movement-related artifacts; Nárai et al. 2022) dataset, which includes T1w images acquired under three different motion conditions. We show that signal-to-noise ratio (SNR) derived metrics are the most important features to predict motion presence.
Scientific Data
Magnetic Resonance Imaging (MRI) provides a unique opportunity to investigate neural changes in healthy and clinical conditions. Its large inherent susceptibility to motion, however, often confounds the measurement. Approaches assessing, correcting, or preventing motion corruption of MRI measurements are under active development, and such efforts can greatly benefit from carefully controlled datasets. We present a unique dataset of structural brain MRI images collected from 148 healthy adults which includes both motion-free and motion-affected data acquired from the same participants. This matched dataset allows direct evaluation of motion artefacts, their impact on derived data, and testing approaches to correct for them. Our dataset further stands out by containing images with different levels of motion artefacts from the same participants, is enriched with expert scoring characterizing the image quality from a clinical point of view and is also complemented with standard image qu...
Human brain mapping, 2017
Motion-contaminated T1-weighted (T1w) magnetic resonance imaging (MRI) results in misestimates of brain structure. Because conventional T1w scans are not collected with direct measures of head motion, a practical alternative is needed to identify potential motion-induced bias in measures of brain anatomy. Head movements during functional MRI (fMRI) scanning of 266 healthy adults (20-89 years) were analyzed to reveal stable features of in-scanner head motion. The magnitude of head motion increased with age and exhibited within-participant stability across different fMRI scans. fMRI head motion was then related to measurements of both quality control (QC) and brain anatomy derived from a T1w structural image from the same scan session. A procedure was adopted to "flag" individuals exhibiting excessive head movement during fMRI or poor T1w quality rating. The flagging procedure reliably reduced the influence of head motion on estimates of gray matter thickness across the cort...
Head motion during MRI acquisition reduces gray matter volume and thickness estimates
NeuroImage, 2015
Imaging biomarkers derived from magnetic resonance imaging (MRI) data are used to quantify normal development, disease, and the effects of disease-modifying therapies. However, motion during image acquisition introduces image artifacts that, in turn, affect derived markers. A systematic effect can be problematic since factors of interest like age, disease, and treatment are often correlated with both a structural change and the amount of head motion in the scanner, confounding the ability to distinguish biology from artifact. Here we evaluate the effect of head motion during image acquisition on morphometric estimates of structures in the human brain using several popular image analysis software packages (FreeSurfer 5.3, VBM8 SPM, and FSL Siena 5.0.7). Within-session repeated T1-weighted MRIs were collected on 12 healthy volunteers while performing different motion tasks, including two still scans. We show that volume and thickness estimates of the cortical gray matter are biased by...
Real‐time 3D motion estimation from undersampled MRI using multi‐resolution neural networks
Medical Physics, 2021
To enable real-time adaptive magnetic resonance imaging-guided radiotherapy (MRIgRT) by obtaining time-resolved three-dimensional (3D) deformation vector fields (DVFs) with high spatiotemporal resolution and low latency (< 500 ms). Theory and Methods:Respiratory-resolved T 1-weighted 4D-MRI of 27 patients with lung cancer were acquired using a golden-angle radial stack-of-stars readout. A multiresolution convolutional neural network (CNN) called TEMPEST was trained on up to 32× retrospectively undersampled MRI of 17 patients, reconstructed with a nonuniform fast Fourier transform, to learn optical flow DVFs. TEMPEST was validated using 4D respiratory-resolved MRI, a digital phantom, and a physical motion phantom. The time-resolved motion estimation was evaluated in-vivo using two volunteer scans, acquired on a hybrid MR-scanner with integrated linear accelerator. Finally, we evaluated the model robustness on a publicly-available four-dimensional computed tomography (4D-CT) dataset. Results: TEMPEST produced accurate DVFs on respiratory-resolved MRI at 20-fold acceleration, with the average end-point-error < 2 mm, both on respiratory-sorted MRI and on a digital phantom. TEMPEST estimated accurate time-resolved DVFs on MRI of a motion phantom, with an error < 2 mm at 28× undersampling. On two volunteer scans, TEMPEST accurately estimated motion compared to the self-navigation signal using 50 spokes per dynamic (366× undersampling).At this undersampling factor,DVFs were estimated within 200 ms, including MRI acquisition. On fully sampled CT data, we achieved a target registration error of 1.87 ± 1.65 mm without retraining the model. Conclusion: A CNN trained on undersampled MRI produced accurate 3D DVFs with high spatiotemporal resolution for MRIgRT.
2020
Visual Abstract This work set out to develop a motion-correction approach aided by conditional generative adversarial network (cGAN) methodology that allows reliable, data-driven determination of involuntary subject motion during dynamic 18F-FDG brain studies. Methods: Ten healthy volunteers (5 men/5 women; mean age ± SD, 27 ± 7 y; weight, 70 ± 10 kg) underwent a test–retest 18F-FDG PET/MRI examination of the brain (n = 20). The imaging protocol consisted of a 60-min PET list-mode acquisition contemporaneously acquired with MRI, including MR navigators and a 3-dimensional time-of-flight MR angiography sequence. Arterial blood samples were collected as a reference standard representing the arterial input function (AIF). Training of the cGAN was performed using 70% of the total datasets (n = 16, randomly chosen), which was corrected for motion using MR navigators. The resulting cGAN mappings (between individual frames and the reference frame [55–60 min after injection]) were then appl...
Scientific Reports
Due to their robustness and speed, recently developed deep learning-based methods have the potential to provide a faster and hence more scalable alternative to more conventional neuroimaging analysis pipelines in terms of whole-brain segmentation based on magnetic resonance (MR) images. These methods were also shown to have higher test–retest reliability, raising the possibility that they could also exhibit superior head motion tolerance. We investigated this by comparing the effect of head motion-induced artifacts in structural MR images on the consistency of segmentation performed by FreeSurfer and recently developed deep learning-based methods to a similar extent. We used state-of-the art neural network models (FastSurferCNN and Kwyk) and developed a new whole-brain segmentation pipeline (ReSeg) to examine whether reliability depends on choice of deep learning method. Structural MRI scans were collected from 110 participants under rest and active head motion and were evaluated fo...
Method for retrospective estimation of natural head movement during structural MRI
Journal of magnetic resonance imaging : JMRI, 2018
Head motion during brain structural MRI scans biases brain morphometry measurements but quantitative retrospective methods estimating head motion from structural MRI have not been evaluated. To verify the hypothesis that two metrics retrospectively computed from MR images: 1) average edge strength (AES, reduced with image blurring) and 2) entropy (ENT, increased with blurring and ringing artifacts) could be sensitive to in-scanner head motion during acquisition of T-weighted MR images. Retrospective. In all, 83 healthy control (HC) and 120 Parkinson's disease (PD) patients. 3D magnetization-prepared rapid gradient-echo (MPRAGE) images at 3T. We 1) compared AES and ENT distribution between HC and PD; 2) evaluated the correlation between tremor score (TS) and AES (or ENT) in PD; and 3) investigated cortical regions showing an association between AES (or ENT) and local and network-level covariance measures of cortical thickness (CT), gray to white matter contrast (GWC) and gray mat...