Improving Quality Control Of MRI Images Using Synthetic Motion Data

Barnaby Nelson; Carrie E. Bearden; Charles Bricout; John Kane; Kang Ik K. Cho; Martha E. Shenton; Michael Harms; Ofer Pasternak; Patrick D. McGorry; Rene S. Kahn

REVIEW 3 major objections 4 minor 26 references

Improving Quality Control Of MRI Images Using Synthetic Motion Data

T0 review · 3 major / 4 minor · reviewed 2026-08-09 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that pretraining an MRI quality-control model on synthetic motion artifacts, with a scalar motion regression objective, beats training from scratch on real unbalanced T1-weighted scans, lifting the rare poor/fair class to…

desk verdict Shows a small but honest transfer gain from synthetic motion regression pretraining for MRI QC; the missing non-motion control keeps the mechanism unproven. read the letter →

arxiv 2502.00160 v2 pith:C7K3VR7F submitted 2025-01-31 eess.IV cs.CV

Charles Bricout , Kang Ik K. Cho , Michael Harms , Ofer Pasternak , Carrie E. Bearden , Patrick D. McGorry , Rene S. Kahn , John Kane

show 5 more authors

Barnaby Nelson Scott W. Woods Martha E. Shenton Sylvain Bouix Samira Ebrahimi Kahou

This is my paper · ORCID

classification eess.IVcs.CV

keywords syntheticmotionartifactMRIqualitycontroltransferlearningclassimbalanceSFCNk-spacecorruptionT1-weighted

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that training a network to estimate a scalar motion score on synthetically corrupted T1-weighted MRI volumes, then transferring the frozen encoder to a quality-control classification task, beats training the same classifier from scratch on real data. On the AMP SCZ dataset, where poor/fair scans are extremely rare, the transferred model reaches a median balanced F1 of 0.13 for the merged poor/fair class and a base accuracy of 0.41, whereas the scratch model scores 0.00 and 0.35. The transfer also uses about 93% less GPU memory, 79% less power, and 94% less training time per use. The practical interest is that a single pretrained motion model could let research teams build their own subjective QC pipelines with very few labeled examples.

What carries the argument

The load-bearing mechanism is the synthetic motion generation pipeline paired with a regression pretraining objective. A k-space corruption step concatenates N randomly sampled affine transforms, with extra elastic deformation, bias field, contrast, flip, and scaling augmentations, and the corresponding ground truth is the RMS deviation of the affine matrices, discretized into 50 bins over [−0.8, 4.8]. Training minimizes the KL divergence between predicted and target bin distributions, following the scalar motion estimation strategy of [9]. For transfer, the SFCN encoder is frozen and a two-layer MLP with batch normalization and ReLU maps its embeddings to the three QC classes.

What would settle it

Retrain the transfer protocol with a control pretraining task on the same synthetic corruption pipeline but with shuffled RMS labels; if the QC gain over scratch is unchanged, then motion-specific regression is not what carries the transfer, and the paper's central explanation would collapse.

Watch

Extended reading notes

Core claim

The central claim is that synthetic motion regression provides a useful inductive bias for downstream QC classification of real brain MRI. The authors synthesize motion by sampling random affine transforms, concatenating their k-space to corrupt volumes, and computing the RMS deviation of the transforms as a continuous ground-truth score; an SFCN encoder is trained to predict a discretized distribution over this score. After training, the encoder weights are frozen and a small multilayer perceptron classifies real AMP SCZ volumes into a three-point quality scale (poor/fair, good, excellent). The paper reports that this transfer protocol outperforms the same classifier trained from scratch on the same real data, especially for the rare poor/fair class, and that the gain is consistent across five random seeds.

Load-bearing premise

The load-bearing premise is that synthetic artifacts created by randomly sampled affine k-space transforms are realistic enough that representations learned to regress their RMS deviation transfer to human QC ratings of real scans; the paper never directly validates this by comparing synthetic artifacts with real motion-corrupted images, and the regression target is generated by the same pipeline that produces the inputs.

Editorial extensions

If this is right

A model pretrained on synthetic motion can be transferred to a new study's subjective QC scale using only 115 labeled training volumes, avoiding the need for large motion-labeled datasets.
Because the transfer task runs in about three minutes on a small GPU, the pretraining cost is amortized after roughly 33 uses, at which point the overall approach becomes more time-efficient than training from scratch.
The same recipe should extend to other artifact types and non-T1w modalities by swapping the synthetic corruption task, though the paper demonstrates it only for 3T T1w data.
A motion-aware encoder pretrained once could serve as a shared backbone across sites and QC scales, reducing the subjectivity of manual QC scoring.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The observed gain might partly reflect generic robustness to augmentation rather than motion-specific knowledge; a control pretraining on the same corruption pipeline with shuffled regression targets would isolate the motion-specific contribution.
A direct test of the pretrained encoder on real scans whose motion was measured prospectively would tell whether the RMS regression features capture genuine motion signatures, a validation the paper does not provide.
A median F1 of 0.13 for the poor/fair class, though an improvement over 0.00, is still low in absolute terms; deployment would require per-site threshold tuning and a larger test set than nine poor/fair volumes.
The reusable asset is the pretrained encoder itself; without broad release, each research team would pay the full pretraining cost, making the 33-use break-even unattractive for single-study users.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Hi X,

Quick take on arXiv:2502.00160: it's a small, honest empirical paper showing that pretraining a motion-score regressor on synthetic MRI artifacts gives a small but consistent boost to downstream QC classification on a heavily unbalanced real dataset. The new bit is using continuous RMS regression as the pretraining objective rather than binary or 5-class classification. It does that carefully: site-level split, five seeds, frozen-encoder transfer compared against a same-architecture scratch model, and the resource table in the paper is transparent about the 25-hour pretraining cost that only pays off after 33 uses.

What it does well: The comparison is fair as far as it goes. Grouping poor/fair into one class is sensible given the data, and the authors admit in the discussion that absolute performance is low. The calibration plot for the motion regressor (R²=0.89) suggests the synthetic pretraining task itself is learned well.

Where it's soft: The biggest hole is the missing non-motion pretraining control. The pretrained model differs from scratch in three ways at once – it has seen ~1000x more volumes, the same synthetic augmentation family, and it's frozen. Without a control that pretrains on, say, a generic corruption regression or shuffled motion labels, we can't tell whether the gain comes from the motion-specific task or just from large-scale synthetic pretraining. That's a real interpretive gap, but it doesn't falsify the paper's literal claim, which is only 'synthetic motion pretraining beats scratch.' The other soft spots are smaller: the key poor/fair class has only 9 test samples and the paper reports medians without any spread, so the 0.13 F1 is fragile. And the abstract's resource claim omits the pretraining cost, though the discussion does the math.

Who it's for: someone working on MRI QC with heavily unbalanced data, or anyone thinking about synthetic pretraining for medical imaging.

My verdict: it deserves serious refereeing. The method is described well enough to reproduce, the comparison is clean, and the missing control is a fixable addition, not a fatal flaw. I'd push for a revision that either adds a non-motion pretraining baseline or narrows the claim to 'synthetic data pretraining helps' rather than 'motion-specific knowledge transfers.'

Referee Report

3 major / 4 minor

Summary. The paper proposes a transfer-learning pipeline for MRI quality control (QC). An SFCN is pretrained on 110,100 synthetically motion-corrupted T1w volumes to regress a discretized RMS motion score, then a small classifier is trained on the frozen encoder's embeddings to predict merged poor/fair, good, and excellent QC classes on real AMP SCZ volumes, using a site-based split. The authors compare this transfer approach with an SFCN trained from scratch on the same downstream training set. They report that the pretrained model achieves a median balanced F1 of 0.13 on the poor/fair class and base accuracy 0.41, versus 0.00 and 0.35 for scratch, and that transfer training requires less GPU memory, power, and time than scratch training. The paper concludes that synthetic motion regression pretraining provides a useful inductive bias for QC classification in unbalanced, data-scarce settings.

Significance. If the observed transfer gain is genuinely attributable to the motion-regression pretraining objective, the approach is a practical contribution to automated MRI QC, where labeled data are scarce and class imbalance is severe. The experimental design has several strengths: a site-based split that prevents site leakage, five random seeds, a frozen-embedding comparison that cleanly isolates the transfer classifier from full fine-tuning, and a fully described synthetic augmentation pipeline. The measured resource savings for the transfer step are concrete. However, the current evidence does not establish that the motion-specific regression objective, rather than generic large-scale synthetic pretraining, is the source of the improvement, and the tiny size of the key test class makes the reported gains statistically fragile. These issues are addressable but are load-bearing for the paper's central claim.

major comments (3)

[Section 3, Table 3; Section 2, Transfer Learning for QC classification] The comparison between the pretrained frozen encoder and the scratch model cannot attribute the improvement to the motion-regression objective, because the two settings differ simultaneously in training data scale (110,100 vs. 115 volumes), exposure to the synthetic augmentation family, and whether the encoder weights are frozen. A control experiment with a non-motion pretraining task, such as regressing a shuffled RMS label or a synthetic non-motion corruption score on the same pipeline, is needed to show that the motion-specific pretraining task, rather than any large-scale synthetic pretraining, provides the inductive bias. As written, the paper's central claim (Discussion, paragraph 1) is undersupported.
[Section 3, Table 3] The results are reported only as medians over five seeds, with no per-seed values or ranges. Because the poor/fair test class contains only 9 samples (Table 2), the difference between median F1 0.13 and 0.00 could be driven by one or two correctly classified volumes. Please report per-seed F1 scores, the full confusion matrices, or confidence intervals so the reader can assess the stability of the claimed gain.
[Abstract; Section 4, Discussion; Table 4] The abstract and introduction claim that the method 'reduces training time and resource requirements compared to training from scratch,' but Table 4 and the Discussion state that if the pretraining cost is included, 33 downstream uses are needed for the overall approach to become more time-efficient. This qualification should appear in the abstract or the results section, and the resource claim should be framed as a break-even analysis over the number of downstream deployments rather than as an unconditional reduction.

minor comments (4)

[Abstract and Introduction] There are typographical errors in the affiliation and introduction: 'Medecine' should be 'Medicine' and 'Scizophrenia' should be 'Schizophrenia'.
[Section 3, Figure 3 caption] The caption reads 'Best R² correlation'; R² is already a squared correlation coefficient, so 'correlation' is redundant. The caption should state what is evaluated on the validation set.
[Section 2, Pretraining for scalar motion estimation] The description of the synthetic pipeline would benefit from stating explicitly whether the affine transforms are sampled from a particular distribution (e.g., uniform ranges on rotation/translation); this is currently only implied by 'randomly apply synthetic motion.'
[References] Reference [20] is a PMLR proceedings paper; please confirm the page numbers or article number are given, since the current citation is incomplete.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the transfer-learning evaluation is an external benchmark on human QC labels, and the synthetic pretraining objective is independently validated on held-out subjects.

full rationale

The paper's derivation chain is self-contained and not circular. The pretraining objective is a regression of the RMS deviation computed from the affine matrices used to synthesize motion (Section 2, 'Pretraining for scalar motion estimation'), and the same section states that 'the ground truth motion score was obtained by computing the RMS deviation from all affine matrices used to generate motion corrupted volumes.' This target is generated by the simulator, but it is an external objective for the pretraining stage, not a quantity derived from the downstream QC labels. The QC evaluation uses real AMP SCZ volumes with manual 4-point scores, splits by site, and compares the pretrained frozen-encoder transfer model with a same-architecture scratch model (Table 3). The reported F1 and accuracy values therefore cannot reduce by construction to the pretraining loss. The binning and KL-divergence loss follow Pollak et al. (reference [9]), which is external prior work, not a self-citation chain, and adopting that modeling choice does not make the result definitional. The paper itself acknowledges the main limitation in the conclusion: 'Even though motion is fully simulated, and our downstream task is different, using transfer learning on a pretrained model yields better results than training from scratch.' That admission concerns the realism of the synthetic-to-real transfer mechanism and the absence of a non-motion pretraining control, which is a confound or generalization gap rather than a circular derivation. No load-bearing step equates a prediction with an input by construction, and no fitted QC parameter is renamed as a prediction.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central claim rests on two unvalidated domain assumptions about the realism and transferability of synthetic motion simulation. The free parameters are standard modeling and tuning choices, none of which are fitted to the downstream QC labels.

free parameters (5)

Regression bin range and count = [-0.8, 4.8] over 50 bins
Hand-chosen discretization of the RMS motion score introduced in Section 2; follows [9] and sets the target distribution the pretraining model learns.
Synthetic samples per volume = 300
Chosen augmentation budget yielding 110100 training, 14100 validation, and 13801 test synthetic volumes; affects pretraining scale.
Transfer classifier hyperparameters = lr=5e-4, dropout=0.7, weight decay=0.05
Tuned via Ray Tune on validation balanced accuracy; reported in Section 2.
Scratch model hyperparameters = lr=3e-6, dropout=0.68, weight decay=0.06
Tuned via Ray Tune for the from-scratch baseline; reported in Section 2.
Synthetic motion augmentation parameters = not stated
The paper says transforms 'randomly apply synthetic motion' with N affine matrices via TorchIO but does not specify the distribution of N, the ranges of translation or rotation, or the elastic and bias field parameters; these choices determine the difficulty of the pretraining task.

assumptions (4)

domain assumption Synthetic k-space affine motion is a realistic proxy for real in-scanner motion.
The entire pretraining pipeline (Section 2) assumes that artifacts generated by composing random affine transforms in k-space resemble real motion artifacts; the paper does not compare generated artifacts to real motion-corrupted images.
domain assumption RMS deviation of the generating affine transforms is a meaningful, transferable summary of motion severity.
The regression target is computed from the synthetic transforms (Section 2); there is no evidence that RMS correlates with human QC ratings on real scans.
standard math KL divergence over a discretized 50-bin distribution is an adequate surrogate for continuous regression.
Standard technique adopted from [9]; unproved but well-established practice.
standard math The k-space concatenation method from [20] produces valid motion-corrupted images.
Adopted as prior art; the paper provides no independent validation.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Improving Quality Control Of MRI Images Using Synthetic Motion Data." pith.science (2026). https://pith.science/paper/C7K3VR7F

@misc{pith2026250200160,
  author       = {Pith},
  title        = {Pith review of: Improving Quality Control Of MRI Images Using Synthetic Motion Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/C7K3VR7F}},
  note         = {Machine review of arXiv:2502.00160}
}

read the original abstract

MRI quality control (QC) is challenging due to unbalanced and limited datasets, as well as subjective scoring, which hinder the development of reliable automated QC systems. To address these issues, we introduce an approach that pretrains a model on synthetically generated motion artifacts before applying transfer learning for QC classification. This method not only improves the accuracy in identifying poor-quality scans but also reduces training time and resource requirements compared to training from scratch. By leveraging synthetic data, we provide a more robust and resource-efficient solution for QC automation in MRI, paving the way for broader adoption in diverse research settings.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 25 canonical work pages

[1]

This modality is how- ever subject to multiple sources of artifacts [1], the most frequent being motion related artifacts

INTRODUCTION Magnetic Resonance Imaging (MRI) has become an invalu- able tool to study the brain anatomy. This modality is how- ever subject to multiple sources of artifacts [1], the most frequent being motion related artifacts. For anatomical T1- weighted (T1w) imaging, even a small amount of motion can bias the estimation of cortical thickness [2]. Whil...

work page
[2]

Improving Quality Control Of MRI Images Using Synthetic Motion Data

MATERIALS AND METHODS Dataset: We use two datasets with manual QC scoring. First, the Human Connectome Project Early Psychosis (HCPEP) includes 390 3T T1w MRI acquired at 3 different sites. Each volume receives a QC score on a 4-point scale (1=poor, 2=fair, 3=good, 4=excellent) [13]. This dataset was used exclusively for synthetic motion generation and tr...

work page Pith review arXiv 2025
[3]

The models performed similarly on the test set, indicating good generalization to unseen data (Figure 3)

RESULTS Pretraining motion estimation task: Our model trained for 25h29m and reached a maximum validation R2 of 0.89. The models performed similarly on the test set, indicating good generalization to unseen data (Figure 3). Fig. 4. Comparison of test accuracy for each transfer model with pretrained task and models trained from scratch QC classification: W...

work page
[4]

Furthermore, models trained to quantify motion appear to learn meaningful embeddings that can be leveraged to per- form QC classification of real MRI data

DISCUSSION Our pretrained model is able to accurately predict the scalar motion score from synthetically motion-corrupted T1w data. Furthermore, models trained to quantify motion appear to learn meaningful embeddings that can be leveraged to per- form QC classification of real MRI data. We also show that training on synthetic data can help with highly unb...

work page
[5]

Available QC datasets are highly unbalanced and use different and subjective scales, making it hard to train one model that fits all scenarios

CONCLUSION MRI research needs automated QC as the number of subjects per study keeps increasing. Available QC datasets are highly unbalanced and use different and subjective scales, making it hard to train one model that fits all scenarios. In this paper, we present an approach to reduce the difficulty of training with unbalanced datasets and in a data-sc...

work page
[6]

Approval was granted by the Research Ethics Committee of ´Ecole de technologie sup´erieure

ETHICS AND ACKNOWLEDGEMENTS This research used retrospective human subject data made available by the NIMH Data Archives. Approval was granted by the Research Ethics Committee of ´Ecole de technologie sup´erieure. We thank the Accelerating Medicines Partner- ship® Schizophrenia (AMP® SCZ), a public-private part- nership managed by the Foundation for the N...

work page
[7]

Imaging artifacts at 3.0T,

M. A. Bernstein et al., “Imaging artifacts at 3.0T,” Journal of Magnetic Resonance Imaging , vol. 24, no. 4, 2006

work page 2006
[8]

Subtle in-scanner motion biases automated measurement of brain anatomy from in vivo MRI,

A. Alexander-Bloch et al., “Subtle in-scanner motion biases automated measurement of brain anatomy from in vivo MRI,” Human Brain Mapping , vol. 37, no. 7, 2016

work page 2016

Show all 26 references

[9]

V olumetric Navigators (vNavs) for Prospective Motion Correction and Selective Reacquisi- tion in Neuroanatomical MRI,

M. D. Tisdall et al., “V olumetric Navigators (vNavs) for Prospective Motion Correction and Selective Reacquisi- tion in Neuroanatomical MRI,” Magnetic Resonance in Medicine, vol. 68, no. 2, pp. 389, Dec. 2011

2011
[10]

Quantifying MR head motion in the Rhineland Study – A robust method for population co- horts,

C. Pollak et al., “Quantifying MR head motion in the Rhineland Study – A robust method for population co- horts,” NeuroImage, vol. 275, pp. 120176, July 2023

2023
[11]

Automated reference-free detec- tion of motion artifacts in magnetic resonance images,

T. K ¨ustner et al., “Automated reference-free detec- tion of motion artifacts in magnetic resonance images,” Magnetic Resonance Materials in Physics, Biology and Medicine, vol. 31, no. 2, pp. 243–256, Apr. 2018

2018
[12]

Motion Artifact Detection for T1- Weighted Brain MR Images Using Convolutional Neu- ral Networks,

E. Roecher et al., “Motion Artifact Detection for T1- Weighted Brain MR Images Using Convolutional Neu- ral Networks,” International Journal of Neural Systems, vol. 34, no. 10, Oct. 2024

2024
[13]

Automatic MR image quality evalua- tion using a Deep CNN: A reference-free method to rate motion artifacts in neuroimaging,

I. Fantini et al., “Automatic MR image quality evalua- tion using a Deep CNN: A reference-free method to rate motion artifacts in neuroimaging,” Computerized Med- ical Imaging and Graphics , vol. 90, pp. 101897, June 2021

2021
[14]

Automatic quality control of brain T1-weighted magnetic resonance images for a clinical data warehouse,

S. Bottani et al., “Automatic quality control of brain T1-weighted magnetic resonance images for a clinical data warehouse,” Medical Image Analysis , vol. 75, pp. 102219, Jan. 2022

2022
[15]

Estimating Head Motion from MR- Images,

C. Pollak et al., “Estimating Head Motion from MR- Images,” Feb. 2023, arXiv:2302.14490 [cs, eess]

2023
[16]

Automatic motion artefact detection in brain T1-weighted magnetic resonance images from a clinical data warehouse using synthetic data,

S. Loizillon et al., “Automatic motion artefact detection in brain T1-weighted magnetic resonance images from a clinical data warehouse using synthetic data,”Medical Image Analysis, vol. 93, pp. 103073, Apr. 2024

2024
[17]

Classifying MRI motion severity using a stacked ensemble approach,

M. Mohebbian et al., “Classifying MRI motion severity using a stacked ensemble approach,” Magnetic Reso- nance Imaging, vol. 75, pp. 107–115, Jan. 2021

2021
[18]

Measuring Transformation Error by RMS Deviation,

M. Jenkinson, “Measuring Transformation Error by RMS Deviation,” Technical TR99MJ1
[19]

An Introduction to the Human Con- nectome Project for Early Psychosis,

G. R Jacobs et al., “An Introduction to the Human Con- nectome Project for Early Psychosis,” Schizophrenia Bulletin, p. sbae123, July 2024

2024
[20]

Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ): Rationale and Study De- sign of the Largest Global Prospective Cohort Study of Clinical High Risk for Psychosis,

C. Wannan et al., “Accelerating Medicines Partnership® Schizophrenia (AMP® SCZ): Rationale and Study De- sign of the Largest Global Prospective Cohort Study of Clinical High Risk for Psychosis,” Schizophrenia Bul- letin, vol. 50, no. 3, pp. 496–512, May 2024

2024
[21]

Clinica: An Open-Source Software Platform for Reproducible Clinical Neuroscience Stud- ies,

A. Routier et al., “Clinica: An Open-Source Software Platform for Reproducible Clinical Neuroscience Stud- ies,” Frontiers in Neuroinformatics, vol. 15, Aug. 2021

2021
[22]

PyTorch: an imperative style, high- performance deep learning library,

A. Paszke et al., “PyTorch: an imperative style, high- performance deep learning library,” in Proceedings of the 33rd International Conference on Neural Informa- tion Processing Systems , number 721, pp. 8026–8037. Curran Associates Inc., Red Hook, NY , USA, Dec. 2019

2019
[23]

PyTorchLightning/pytorch-lightning: 0.7.6 release,

W. Falcon et al., “PyTorchLightning/pytorch-lightning: 0.7.6 release,” May 2020

2020
[24]

MONAI: An open-source framework for deep learning in healthcare,

M. J. Cardoso et al., “MONAI: An open-source framework for deep learning in healthcare,” Nov. 2022, arXiv:2211.02701 [cs]

2022 arXiv
[25]

TorchIO: A Python library for ef- ficient loading, preprocessing, augmentation and patch- based sampling of medical images in deep learning,

F. P ´erez-Garc´ıa et al., “TorchIO: A Python library for ef- ficient loading, preprocessing, augmentation and patch- based sampling of medical images in deep learning,” Computer Methods and Programs in Biomedicine , vol. 208, pp. 106236, Sept. 2021

2021
[26]

MRI k-Space Motion Artefact Aug- mentation: Model Robustness and Task-Specific Uncer- tainty,

R. Shaw et al., “MRI k-Space Motion Artefact Aug- mentation: Model Robustness and Task-Specific Uncer- tainty,” in Proceedings of The 2nd International Con- ference on Medical Imaging with Deep Learning . May 2019, PMLR. 4

2019

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.

[1] [1]

This modality is how- ever subject to multiple sources of artifacts [1], the most frequent being motion related artifacts

INTRODUCTION Magnetic Resonance Imaging (MRI) has become an invalu- able tool to study the brain anatomy. This modality is how- ever subject to multiple sources of artifacts [1], the most frequent being motion related artifacts. For anatomical T1- weighted (T1w) imaging, even a small amount of motion can bias the estimation of cortical thickness [2]. Whil...

work page

[2] [2]

Improving Quality Control Of MRI Images Using Synthetic Motion Data

MATERIALS AND METHODS Dataset: We use two datasets with manual QC scoring. First, the Human Connectome Project Early Psychosis (HCPEP) includes 390 3T T1w MRI acquired at 3 different sites. Each volume receives a QC score on a 4-point scale (1=poor, 2=fair, 3=good, 4=excellent) [13]. This dataset was used exclusively for synthetic motion generation and tr...

work page Pith review arXiv 2025

[3] [3]

The models performed similarly on the test set, indicating good generalization to unseen data (Figure 3)

RESULTS Pretraining motion estimation task: Our model trained for 25h29m and reached a maximum validation R2 of 0.89. The models performed similarly on the test set, indicating good generalization to unseen data (Figure 3). Fig. 4. Comparison of test accuracy for each transfer model with pretrained task and models trained from scratch QC classification: W...

work page

[4] [4]

Furthermore, models trained to quantify motion appear to learn meaningful embeddings that can be leveraged to per- form QC classification of real MRI data

DISCUSSION Our pretrained model is able to accurately predict the scalar motion score from synthetically motion-corrupted T1w data. Furthermore, models trained to quantify motion appear to learn meaningful embeddings that can be leveraged to per- form QC classification of real MRI data. We also show that training on synthetic data can help with highly unb...

work page

[5] [5]

Available QC datasets are highly unbalanced and use different and subjective scales, making it hard to train one model that fits all scenarios

CONCLUSION MRI research needs automated QC as the number of subjects per study keeps increasing. Available QC datasets are highly unbalanced and use different and subjective scales, making it hard to train one model that fits all scenarios. In this paper, we present an approach to reduce the difficulty of training with unbalanced datasets and in a data-sc...

work page

[6] [6]

Approval was granted by the Research Ethics Committee of ´Ecole de technologie sup´erieure

ETHICS AND ACKNOWLEDGEMENTS This research used retrospective human subject data made available by the NIMH Data Archives. Approval was granted by the Research Ethics Committee of ´Ecole de technologie sup´erieure. We thank the Accelerating Medicines Partner- ship® Schizophrenia (AMP® SCZ), a public-private part- nership managed by the Foundation for the N...

work page

[7] [7]

Imaging artifacts at 3.0T,

M. A. Bernstein et al., “Imaging artifacts at 3.0T,” Journal of Magnetic Resonance Imaging , vol. 24, no. 4, 2006

work page 2006

[8] [8]

Subtle in-scanner motion biases automated measurement of brain anatomy from in vivo MRI,

A. Alexander-Bloch et al., “Subtle in-scanner motion biases automated measurement of brain anatomy from in vivo MRI,” Human Brain Mapping , vol. 37, no. 7, 2016

work page 2016