Pith. sign in

REVIEW 3 major objections 8 minor 50 references

DISARM++: Beyond scanner-free harmonization

T0 review · 3 major / 8 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read DISARM++ removes scanner bias from 3D MRI without skull-stripping and beats two benchmark harmonizers on every test.

desk verdict Harmonization evidence is strong; downstream validation and the scanner-free identity need more work before the SOTA claims hold. read the letter →

arxiv 2505.03715 v1 pith:LZBEFA2T submitted 2025-05-06 cs.CV

Luca Caldera , Lara Cavinato , Alessio Cirone , Isabella Cama , Sara Garbarino , Raffaele Lodi , Fabrizio Tagliavini , Anna Nigri
show 4 more authors
Silvia De Francesco Andrea Cappozzo Michele Piana Francesca Ieva
This is my paper · ORCID
classification cs.CV
keywords imageharmonizationimage-to-imagetranslationmagneticresonanceimagingscanner-freedisentangledrepresentationsbrainagepredictionAlzheimer'sdiseaseclassificationmulti-siteMRI
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 proposes that scanner effects in T1-weighted MR images can be removed at the image level rather than at the feature level, by teaching a network to separate what a brain looks like from which scanner acquired it. Its model, DISARM++, can send any scan into a scanner-free reference space, or restyle it to match a chosen training scanner, working on full-head volumes with no skull-stripping and no retraining for new scanner types. The authors test it on healthy controls, traveling subjects scanned on multiple machines, and Alzheimer's disease patients, and report that it outperforms two image-based state-of-the-art methods across harmonization quality, age prediction, inter-scanner volume variability, Alzheimer's classification, and diagnosis prediction. If these results hold, the contribution is a drop-in harmonization step that makes multi-site neuroimaging data comparable while preserving the whole head for downstream analysis.

What carries the argument

The argument is carried by a disentangled latent-space design. A brain encoder $E_b$ maps an image to an anatomical-content vector $z^b$, a variational scanner encoder $E_s$ maps the same image and its scanner label to a scanner-effect distribution whose sampled vector $z^s_i = \sigma_i \epsilon + \mu_i$ encodes the scanner style, and a generator $G$ recomposes the image from both. Training swaps the anatomical vectors between two scanner domains to enforce cycle-consistency, while a new scanner-free loss $L_{\mathrm{sf}}$ demands that images generated from the same Gaussian noise through different anatomies map to the same scanner encoding, which is what makes the scanner-free output act like a denoised, scanner-independent image. Attention layers in the encoders and generator, together with a 26-slice moving window, are the architectural changes that let the model retain anatomical detail while processing thinner volumes.

What would settle it

On a traveling-subjects set where the same person is scanned on several machines, compute the pairwise similarity of scanner-free outputs across scanners and compare each output against the person's known anatomy, such as a real lesion. The claim fails if scanner-free outputs from the same subject remain scanner-dependent, or if two different subjects' scanner-free outputs become more alike than their raw scans, or if a visible lesion is erased in the scanner-free version.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims that a single generator can produce a scanner-free version of any T1-weighted brain scan by discarding the scanner-specific component of the image and regenerating the anatomy from latent content plus Gaussian noise. The same machinery also allows the image to be transplanted into any training-scanner domain. Across healthy controls, traveling subjects, and Alzheimer's patients, the harmonized images come out visually consistent, their voxel-intensity distributions converge, and downstream models perform better; age prediction reaches $R^2 \approx 0.60$, Alzheimer's versus healthy classification accuracy reaches $0.86$, and diagnosis of mild cognitive impairment versus Alzheimer's disease reaches an AUC of $0.95$, in each case above the two benchmark harmonizers. The paper further claims this works for scanners never seen during training and without skull-stripping, which it treats as an advantage for full-head applications.

Load-bearing premise

The model assumes an MR image splits cleanly into anatomy and scanner effect, so that replacing the scanner part with Gaussian noise leaves all biologically relevant signal untouched.

Editorial extensions

If this is right

  • Multi-site MRI studies can be pooled without feature-level harmonization or skull-stripping, because the harmonization happens on the image itself.
  • Scans from new scanners can be harmonized by inference only, with no retraining or fine-tuning, which lowers the barrier to adding new sites to a study.
  • Two harmonization modes are available: scanner-free output for general pooling, and reference-scanner style transfer when a downstream model was trained on one specific scanner's look.
  • Because the whole head is preserved, the method can support analyses outside brain tissue, such as head trauma and cranial deformation.
  • Downstream predictive tasks such as age estimation, Alzheimer's screening, and diagnosis inherit the harmonization benefit directly, rather than requiring a separate feature-level correction.

Reading between the lines

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

  • If the disentanglement holds, the same scanner-free space could serve as a common substrate for cross-site federated learning, where models share harmonized images or features without sharing raw data; the paper does not discuss this.
  • A testable extension is to apply DISARM++ to a paired traveling-subject dataset with a visible anatomical abnormality; the scanner-free outputs should agree across scanners while still showing the abnormality, directly testing whether pathology is preserved.
  • The scanner-free encoding could also be used as a normalization step for image retrieval or for training generative models on pooled multi-site data, though the paper only evaluates the four downstream tasks.
  • Replacing the scanner code with Gaussian noise may also suppress scanner-specific image fingerprints, which raises a possible anonymization use that the paper does not claim.
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.

Referee Report

3 major / 8 minor

Summary. This paper proposes DISARM++, an unsupervised image-to-image translation model for harmonizing 3D T1-weighted MRI. The model assumes each image is generated from a brain-structure latent B and a scanner-effect latent S; after training with cycle-consistency, adversarial, classification, KL, and a new scanner-free loss, inference maps a new image either to one of the training scanners or to a 'scanner-free' space by substituting random Gaussian noise for the scanner effect. The authors train on 701 healthy-control images from five scanners and evaluate on ten scanners (including scanners unseen in training), on traveling subjects, and on four downstream tasks: age prediction, inter-scanner volume variability, AD-versus-healthy classification, and diagnosis prediction. They report strong harmonization improvements over STGAN and IGUANe (e.g., pairwise JSD dropping from about 0.17 to 0.009), better traveling-subject SSIM, and better downstream results (age R2≈0.60, AD accuracy≈0.86, diagnosis AUC≈0.95).

Significance. The practical goal is important: direct voxel-level harmonization that does not require skull-stripping and generalizes to unseen scanners would be useful for multi-center neuroimaging. The paper's strengths include a broad evaluation across multiple public and private datasets, a traveling-subject test, comparison with two strong baselines, and several downstream tasks; this is a substantial amount of evidence for the core harmonization effect. The traveling-subject SSIM improvements and the reduction in inter-scanner volume variability (Table 9) are the most convincing parts of the empirical package. However, the 'scanner-free' claim is not independently established by the current metrics, and the headline diagnosis AUC appears to be an in-sample estimate. The paper is therefore promising but needs additional validation before the central claims can be accepted.

major comments (3)
  1. [Section 3.4, Eq. (7)] The 'scanner-free' property is the load-bearing claim of the paper, but the scanner-free loss in Eq. (7) only enforces Es(hat{x}_{k->f}) ≈ Es(hat{x}_{h->f}) for two outputs generated from the same Gaussian noise epsilon; it does not anchor these latents to the prior N(0,1) or to any canonical reference appearance. Nothing in Eq. (8) as described applies L_KL or L_lat to the scanner-free encodings, so the generator could map all inputs to a new, arbitrary common domain F that is not 'scanner-free' in any biologically meaningful sense. The JSD/HD/WD reductions in Table 6 and the AD-test acceptance only show that the ten scanner-mean intensity distributions are closer to each other after harmonization; they do not distinguish the claimed scanner-free space from a learned blend. Please add a direct probe of scanner information remaining in scanner-free outputs (e.g., accuracy of a scanner classifier trained on original images and tested on scanner-free images), clarify exactly which losses are applied to Es(hat{x}_{->f}), and show that scanner-free outputs from the same brain but different source scanners agree at the voxel level beyond the traveling-subject SSIM already reported.
  2. [Section 5.4.4 / Section 6.3.4] The diagnosis-prediction result is the clearest over-claim. With 41 subjects and a PCA plus logistic-regression pipeline, the single AUC of 0.9459 reported in Section 6.3.4 is almost certainly an in-sample estimate: Section 5.4.4 describes PCA and logistic regression but no cross-validation or held-out test set. An in-sample AUC on 41 subjects can be severely optimistic and cannot be compared with the other methods' AUCs, which appear to be computed in the same way. The authors should re-estimate the AUC with repeated stratified cross-validation, with PCA fitted inside each training fold, and report the distribution over folds plus a permutation null.
  3. [Section 5.4.3 / Table 10] The AD-versus-healthy classification is confounded: healthy images come from RIN, IXI, and PPMI, while AD images come from NeuroArtP3 and ADNI3 (Section 5.4.3). Scanner and dataset are therefore partially predictive of the label even before any biology is considered. The improved accuracy after DISARM++ harmonization (0.858 ± 0.03) could reflect removal of some scanner effects, but it does not by itself show preservation of disease-related signal. Please provide scanner-balanced cross-validation folds or report a scanner-label classification accuracy on the same harmonized images as a sanity check.
minor comments (8)
  1. [Section 5.4.2] The text says 'yij and xij represent the age and volume, respectively', while the model y = β0 + β1 x + u + ε predicts volume from age; the wording appears to be reversed and should be corrected.
  2. [Section 5.4.3 / Figure 12] The sentence 'DISARM++ shows significantly better precision, F1 score, and recall, although no significant difference in precision is observed' is self-contradictory; the first occurrence of 'precision' should presumably be 'accuracy' or the sentence should be rephrased.
  3. [Table 5] The configuration without Lsf achieves a lower post-harmonization JSD (0.004) than the full model (0.008); the statement that all components are 'crucial' for harmonization is therefore only supported jointly with the structural metrics, and should be phrased accordingly.
  4. [Section 3.2 / Eq. (4)] There is a notation inconsistency in the method description: Section 3.2 defines the brain discriminator as Db : X → C, processing images, but Eq. (4) evaluates Db on the brain-structure latent zb. If the discriminator actually operates on the latent space B, the architecture description and Figure 1 should be corrected; if it operates on images, Eq. (4) is wrong.
  5. [Figures 3 and 5 / Section 3.1] The vector c0 used for scanner-free generation is not defined in Section 3.1, where C is defined as one-hot vectors with ∥c∥1 = 1; since c0 appears to be neither one-hot nor a member of C, its meaning should be stated explicitly.
  6. [Section 5.4.1] The age prediction comparison reports R2 and RMSE from 10-fold cross-validation but no paired significance test; given the overlap between the DISARM++ and raw-image standard deviations, a paired test would strengthen the claim of superiority.
  7. [General] Several typos remain, e.g., 'distict' (Section 5.3.2), 'Hellringer' (Appendix D), and 'F ormula' in Table 2; also, the code link is given only as 'this link' and should be a permanent URL.
  8. [Section 3.5] Inference for scanner-free harmonization uses random Gaussian noise epsilon; the paper should specify whether the same epsilon is used when harmonizing two scans of the same subject from different scanners, otherwise part of the traveling-subject SSIM comparison may be affected by stochasticity.

Circularity Check

1 steps flagged · score 6.0 of 10

Diagnosis AUC is an in-sample logistic-regression fit; the remaining harmonization and downstream claims are not circular.

  1. fitted input called prediction [Section 5.4.4 (Diagnosis Prediction) and Section 6.3.4]
    "To reduce dimensionality, we applied principal component analysis (PCA) and retained the number of components that explain 70% of the variance... We then use the principal components (PCs) for logistic regression modeling of the classification task. For performance evaluation, we report the area under the ROC curve (AUC). ... After applying DISARM++ harmonization, the AUC significantly improves to 0.9459."

    Section 5.4.4 describes no held-out split, cross-validation, or bootstrap for the 41 NeuroArtP3 subjects: PCA and logistic regression are fit on the same 41 subjects whose AUC is then reported. The reported AUC therefore measures in-sample discrimination of the fitted model, not an out-of-sample prediction. With 57 volume variables reduced by PCA on n=41, the improvement from raw AUC 0.6284 to 0.9459 is the expected result of fitting and evaluating on identical data, so the headline 'diagnosis prediction' statistic is a fitted number renamed as a prediction.

full rationale

The core image harmonization derivation is not circular: the disentanglement assumption is stated explicitly in Section 3.1, the losses in Section 3.4 are defined independently of the evaluation metrics, and the main harmonization claims are tested on unseen scanners, traveling subjects, and held-out downstream tasks. Age prediction uses 10-fold cross-validation (Section 5.4.1), AD classification uses 10 random splits (Section 5.4.3), and the traveling-subject SSIM analysis is a genuine external check. The harmonization metrics JSD/HD/WD overlap with the training objective, but they are computed on held-out scanner distributions and are not identical to any loss by construction, so this is objective alignment rather than circularity. The DISARM baseline is a self-citation, but it is only an ablation baseline and is not load-bearing for the new scanner-free loss or the main claims. The single genuine circular step is the diagnosis prediction AUC (0.9459), which is an in-sample logistic-regression fit presented as a predictive result; the manuscript provides no validation protocol for that task. Because one headline downstream number reduces by construction to its own fit, the paper is partially circular, warranting score 6.

Assumptions & free parameters 4 free parameters · 4 assumptions · 3 invented entities

The central method rests on a small number of modeling assumptions and hand-set hyperparameters. The key burden is the disentanglement assumption: if scanner and anatomy are not separable, the scanner-free output is not trustworthy. The paper also chooses loss weights, latent dimensions, a PCA threshold, and a sliding window width without sensitivity analyses.

free parameters (4)
  • Loss weights (lambda_cc, lambda_rec, lambda_lat, lambda_KL, lambda_sf, lambda_b_adv, lambda_s_cls, lambda_s_adv) = 10, 10, 8, 0.01, 7, 1, 3, 10
    Hand-chosen values reported in Appendix B Table .14; no sensitivity analysis is provided, only whole-component removal in the ablation.
  • Latent dimensions dim(B), dim(S) = (91, 109, 91) and 16
    Architectural hyperparameters set by hand; no ablation or justification for these sizes is given.
  • PCA variance threshold for diagnosis model = 70%
    The number of retained components in Section 5.4.4 is chosen as the threshold explaining 70% of variance; the threshold is not varied and no validation is described.
  • Partial volume sliding window width = 26 slices
    Introduced to reduce memory and allow deeper layers; no comparison of window widths is reported.
assumptions (4)
  • domain assumption An MR image x can be written as a combination of independent anatomical content B and scanner effect S such that swapping B between scanners preserves all biological information.
    Section 3.1 assumes this disentanglement; if scanner and anatomy are entangled, the scanner-free output will not be biologically faithful.
  • domain assumption Scanner differences in T1-weighted MRI act as a transferable style that can be modeled by a generator conditioned on a scanner label.
    Underlies the I2I architecture inherited from DRIT++; nonlinear scanner-image interactions would break the label conditioning.
  • domain assumption Cycle-consistency and self-reconstruction losses are sufficient to guarantee anatomical content preservation.
    Relied on in Section 3.4; cycle-consistency is a heuristic, not a proof of content preservation.
  • domain assumption Traveling subjects scanned at different sites provide an identity benchmark: harmonized images of the same subject should be similar.
    Used in Sections 5.3.2 and 6.2.3; assumes no physiological or positional confounds between sessions beyond scanner effects.
invented entities (3)
  • Scanner-free space F
    purpose: Target output domain for harmonized images, produced by the generator with random Gaussian noise and a null scanner label c0.
    F is defined only by the model's outputs; no external standard verifies that images in F are scanner-free beyond distributional similarity and downstream metrics.
  • Scanner effect latent space S
    purpose: Stores scanner-specific style as a VAE mean and variance per scanner domain.
    Internal representation; no independent measurement separates S from anatomical content.
  • Brain structure latent space B
    purpose: Supposed scanner-invariant encoding of anatomy used for cross-scanner swapping.
    Internal representation; no direct evidence that B is fully scanner-invariant.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DISARM++: Beyond scanner-free harmonization." pith.science (2026). https://pith.science/paper/LZBEFA2T

@misc{pith2026250503715,
  author       = {Pith},
  title        = {Pith review of: DISARM++: Beyond scanner-free harmonization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LZBEFA2T}},
  note         = {Machine review of arXiv:2505.03715}
}
read the original abstract

Harmonization of T1-weighted MR images across different scanners is crucial for ensuring consistency in neuroimaging studies. This study introduces a novel approach to direct image harmonization, moving beyond feature standardization to ensure that extracted features remain inherently reliable for downstream analysis. Our method enables image transfer in two ways: (1) mapping images to a scanner-free space for uniform appearance across all scanners, and (2) transforming images into the domain of a specific scanner used in model training, embedding its unique characteristics. Our approach presents strong generalization capability, even for unseen scanners not included in the training phase. We validated our method using MR images from diverse cohorts, including healthy controls, traveling subjects, and individuals with Alzheimer's disease (AD). The model's effectiveness is tested in multiple applications, such as brain age prediction (R2 = 0.60 \pm 0.05), biomarker extraction, AD classification (Test Accuracy = 0.86 \pm 0.03), and diagnosis prediction (AUC = 0.95). In all cases, our harmonization technique outperforms state-of-the-art methods, showing improvements in both reliability and predictive accuracy. Moreover, our approach eliminates the need for extensive preprocessing steps, such as skull-stripping, which can introduce errors by misclassifying brain and non-brain structures. This makes our method particularly suitable for applications that require full-head analysis, including research on head trauma and cranial deformities. Additionally, our harmonization model does not require retraining for new datasets, allowing smooth integration into various neuroimaging workflows. By ensuring scanner-invariant image quality, our approach provides a robust and efficient solution for improving neuroimaging studies across diverse settings. The code is available at this link.

Figures

Figures reproduced from arXiv: 2505.03715 by the authors.

Figure 1
Figure 1. From left to right, we have a) the brain encoder [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. A high-level functional diagram of the model training procedure. The encoders [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. During each iteration, the training procedure involves randomly [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (9 more)
Figure 3
Figure 3. Figure 3: Training procedure part related to the scanner-free loss. The images used to illustrate the procedure are acquired with the Prisma (purple) and Gyroscan Intera (green) scanners. Moreover, as shown in [PITH_FULL_IMAGE:figures/full_fig_p009_3.png]
Figure 4
Figure 4. Figure 4: Inference to a reference scanner for a new image [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]
Figure 5
Figure 5. Figure 5: Inference to scanner-free for a new image xj ∈ Xj . Specifically, the figure illustrates the transfer of an image acquired with the Skyra Fit (pink) scanner to the scanner-free space (black). 4. Datasets and Preprocessing 4.1. Datasets We obtained T1-weighted MR images…
Figure 6
Figure 6. Figure 6: DISARM++ harmonization visual assessment. The figure displays slices from [PITH_FULL_IMAGE:figures/full_fig_p026_6.png]
Figure 7
Figure 7. Figure 7: IGUANe harmonization visual assessment. The figure displays slices from the [PITH_FULL_IMAGE:figures/full_fig_p027_7.png]
Figure 8
Figure 8. Figure 8: STGAN harmonization visual assessment. The figure displays slices from the [PITH_FULL_IMAGE:figures/full_fig_p028_8.png]
Figure 9
Figure 9. Figure 9: Comparison of mean voxel intensity distributions from the 10 test scanners [PITH_FULL_IMAGE:figures/full_fig_p031_9.png]
Figure 10
Figure 10. Figure 10: For subject 1, we display a sagittal slice from three MR images [PITH_FULL_IMAGE:figures/full_fig_p032_10.png]
Figure 10
Figure 10. Figure 10: Visual harmonization comparison for traveling subjects. For Subject 6, we [PITH_FULL_IMAGE:figures/full_fig_p034_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 46 canonical work pages

  1. [1]

    Alotaibi, A. (2020). Deep generative adversarial networks for image-to- image translation: A review. Symmetry, 12(10):1705

    work page 2020

  2. [2]

    H., Yeo, B

    An, L., Chen, J., Chen, P., Zhang, C., He, T., Chen, C., Zhou, J. H., Yeo, B. T., of Aging, L. S., Initiative, A. D. N., et al. (2022). Goal-specific brain mri harmonization. Neuroimage, 263:119570

    work page 2022

  3. [3]

    Brain-developement ixi

    Brain-Developement (2019). Brain-developement ixi. https:// brain-development.org/ixi-dataset/

    work page 2019

  4. [4]

    Caldera, L., Cavinato, L., Cappozzo, A., Cama, I., Garbarino, S., Cirone, A., Lodi, R., Tagliavini, F., Nigri, A., De Francesco, S., et al. (2025). Disarm: Disentangled scanner-free image generation via unsupervised im- age2image translation. In International Workshop on Machine Learning in Clinical Neuroimaging, pages 102–112. Springer

    work page 2025

  5. [5]

    C., Sollini, M., Kirienko, M., and Ieva, F

    Cavinato, L., Massi, M. C., Sollini, M., Kirienko, M., and Ieva, F. (2023). Dual adversarial deconfounding autoencoder for joint batch-effects removal from multi-center and multi-scanner radiomics data. Scientific Reports, 13(1):18857

    work page 2023

  6. [6]

    A., Beer, J

    Chen, A. A., Beer, J. C., Tustison, N. J., Cook, P. A., Shinohara, R. T., Shou, H., and Initiative, A. D. N. (2022). Mitigating site effects in covari- ance for machine learning in neuroimaging data. Human brain mapping , 43(4):1179–1195

    work page 2022

  7. [7]

    Choi, Y., Uh, Y., Yoo, J., and Ha, J.-W. (2020). Stargan v2: Diverse image synthesis for multiple domains. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 8188–8197

    work page 2020

  8. [8]

    H., Busa, E., Albert, M., Dieterich, M., Haselgrove, C., van der Kouwe, A., Killiany, R., Kennedy, D., Klaveness, S., Montillo, A., Makris, N., Rosen, B., and Dale, A

    Fischl, B., Salat, D. H., Busa, E., Albert, M., Dieterich, M., Haselgrove, C., van der Kouwe, A., Killiany, R., Kennedy, D., Klaveness, S., Montillo, A., Makris, N., Rosen, B., and Dale, A. M. (2002). Whole brain segmen- tation: automated labeling of neuroanatomical structures in the human brain. Neuron, 33:341–355

    work page 2002

Show all 50 references
  1. [9]

    S., Evans, A

    Fonov, V. S., Evans, A. C., McKinstry, R. C., Almli, C. R., and Collins, D. (2009). Unbiased nonlinear average age-appropriate brain templates from birth to adulthood. NeuroImage, 47:S102. 57

    2009

  2. [10]

    Guo, Z., Gu, Z., Zheng, B., Dong, J., and Zheng, H. (2022). Transformer for image harmonization and beyond. IEEE Transactions on Pattern Anal- ysis and Machine Intelligence

    2022

  3. [11]

    Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., and Hochreiter, S. (2017). Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems , 30

    2017

  4. [12]

    Isensee, F., Schell, M., Pflueger, I., Brugnara, G., Bonekamp, D., Neu- berger, U., Wick, A., Schlemmer, H.-P., Heiland, S., Wick, W., Bendszus, M., H Maier-Hein, K., and Kickingereder, P. (2019). Automated brain extraction of multisequence mri using artificial neural networks...

    2019

  5. [13]

    R., Bernstein, M

    Jack Jr, C. R., Bernstein, M. A., Fox, N. C., Thompson, P., Alexander, G., Harvey, D., Borowski, B., Britson, P. J., L. Whitwell, J., Ward, C., Dale, A. M., Felmlee, J. P., Gunter, J. L., Hill, D. L. G., Killiany, R., Schuff, N., Fox-Bosetti, S., Lin, C., Studholme, C., DeCarl...

    2008

  6. [14]

    Jenkinson, M., Bannister, P., Brady, M., and Smith, S. (2002). Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage, 17(2):825–841

    2002

  7. [15]

    F., Behrens, T

    Jenkinson, M., Beckmann, C. F., Behrens, T. E., Woolrich, M. W., and Smith, S. M. (2012). Fsl. Neuroimage, 62(2):782–790

    2012

  8. [16]

    L., Archibald, S

    Jernigan, T. L., Archibald, S. L., Fennema-Notestine, C., Gamst, A. C., Stout, J. C., Bonner, J., and Hesselink, J. R. (2001). Effects of age on tissues and regions of the cerebrum and cerebellum. Neurobiology of aging, 22(4):581–594

    2001

  9. [17]

    Kailath, T. (2003). The divergence and bhattacharyya distance mea- sures in signal selection. IEEE transactions on communication technology, 15(1):52–60. 58

    2003

  10. [18]

    Lee, H.-Y., Tseng, H.-Y., Mao, Q., Huang, J.-B., Lu, Y.-D., Singh, M., and Yang, M.-H. (2020). Drit++: Diverse image-to-image translation via disentangled representations. International Journal of Computer Vision , 128:2402–2417

    2020

  11. [19]

    Lin, J. (1991). Divergence measures based on the shannon entropy. IEEE Transactions on Information theory , 37(1):145–151

    1991

  12. [20]

    H., Maiti, P., Thomopoulos, S

    Liu, M., Zhu, A. H., Maiti, P., Thomopoulos, S. I., Gadewar, S., Chai, Y., Kim, H., Jahanshad, N., and Initiative, A. D. N. (2023). Style transfer generative adversarial networks to harmonize multisite mri to a single ref- erence image to avoid overcorrection.Human Brain Mappi...

    2023

  13. [21]

    and Yap, P.-T

    Liu, S. and Yap, P.-T. (2024). Learning multi-site harmonization of magnetic resonance images without traveling human phantoms. Commu- nications Engineering, 3(1):6

    2024

  14. [22]

    Malaguti, M. C., Gios, L., Giometto, B., Longo, C., Riello, M., Ot- taviani, D., Pellegrini, M., Di Giacopo, R., Donner, D., Rozzanigo, U., Chierici, M., Moroni, M., Jurman, G., Bincoletto, G., Pardini, M., Bacchin, R., Nobili, F., Di Biasio, F., Avanzino, L., Marchese, R., Ma...

    2024

  15. [23]

    L., Ostrowizki, S., Fontoura, P., La-Roche, F

    Marek, K., Jennings, D., Lasch, S., Siderowf, A., Tanner, C., Simuni, T., Coffey, C., Kieburtz, K., Flagg, E., Chowdhury, S., Poewe, W., Mollen- hauer, B., Sherer, T., Frasier, M., Meunier, C., Rudolph, A., Casaceli, C., Seibyl, J., Mendick, S., Schuff, N., Zhang, Y., Toga, A....

    2011

  16. [24]

    M., and Thompson, P

    Moyer, D., Ver Steeg, G., Tax, C. M., and Thompson, P. M. (2020). Scanner invariant representations for diffusion mri harmonization. Mag- netic resonance in medicine , 84(4):2174–2189

    2020

  17. [25]

    Nigri, A., Ferraro, S., Wheeler-Kingshott, C. A. G., Tosetti, M., Redolfi, A., Forloni, G., D’Angelo, E., Aquino, D., Biagi, L., Bosco, P., Carne, I., De Francesco, S., Demichelis, G., Gianeri, R., Lagana, M. M., Micotti, E., Napolitano, A., Palesi, F., Pirastru, A., Savini, G...

    2022

  18. [26]

    Nikulin, M. S. (2001). Hellinger distance. In Hazewinkel, M., editor, Encyclopedia of Mathematics . Springer. https://encyclopediaofmath. org/wiki/Hellinger_distance

    2001

  19. [27]

    Panaretos, V. M. and Zemel, Y. (2019). Statistical aspects of wasserstein distances. Annual review of statistics and its application , 6(1):405–431

    2019

  20. [28]

    P´ erez-Garc´ ıa, F., Sparks, R., and Ourselin, S. (2021). Torchio: a python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning. Computer Methods and Pro- grams in Biomedicine, 208:106236

    2021

  21. [29]

    M., Satterthwaite, T

    Pomponio, R., Erus, G., Habes, M., Doshi, J., Srinivasan, D., Mamourian, E., Bashyam, V., Nasrallah, I. M., Satterthwaite, T. D., Fan, Y., et al. (2020). Harmonization of large mri datasets for the analysis of brain imaging patterns throughout the lifespan. NeuroImage, 208:116450. 60

    2020

  22. [30]

    J., Weickert, C

    Radua, J., Vieta, E., Shinohara, R., Kochunov, P., Quid´ e, Y., Green, M. J., Weickert, C. S., Weickert, T., Bruggemann, J., Kircher, T., Ne- nadi´ c, I., Cairns, M. J., Seal, M., Schall, U., Henskens, F., Fullerton, J. M., Mowry, B., Pantelis, C., Lenroot, R., Cropley, V., Lo...

    2020

  23. [31]

    Rao, C. R. (1995). A review of canonical coordinates and an alternative to correspondence analysis using hellinger distance. Q¨ uestii´ o: quaderns d’estad´ ıstica i investigaci´ o operativa

    1995

  24. [32]

    Roca, V., Kuchcinski, G., Pruvo, J.-P., Manouvriez, D., and Lopes, R. (2024). Iguane: a 3d generalizable cyclegan for multicenter harmonization of brain mr images. arXiv preprint arXiv:2402.03227

    2024 arXiv

  25. [33]

    Scholz, F. W. and Stephens, M. A. (1987). K-sample anderson–darling tests. Journal of the American Statistical Association , 82(399):918–924

    1987

  26. [34]

    M., Busa, E., Glessner, M., Salat, D., Hahn, H

    S´ egonne, F., Dale, A. M., Busa, E., Glessner, M., Salat, D., Hahn, H. K., and Fischl, B. (2004). A hybrid approach to the skull stripping problem in mri. Neuroimage, 22(3):1060–1075

    2004

  27. [35]

    Sele, S., Liem, F., M´ erillat, S., and J¨ ancke, L. (2020). Decline variability of cortical and subcortical regions in aging: A longitudinal study. Frontiers in human neuroscience, 14:363

    2020

  28. [36]

    T., Oh, J., Nair, G., Calabresi, P

    Shinohara, R. T., Oh, J., Nair, G., Calabresi, P. A., Davatzikos, C., Doshi, J., Henry, R. G., Kim, G., Linn, K. A., Papinutto, N., Pelletier, D., 61 Pham, D. L., Reich, D. S., Rooney, W., Roy, S., Stern, W., Tummala, S., Yousuf, F., Zhu, A., Sicotte, N. L., Bakshi, R., and Co...

    2017

  29. [37]

    G., Zijdenbos, A

    Sled, J. G., Zijdenbos, A. P., and Evans, A. C. (1998). A nonparametric method for automatic correction of intensity nonuniformity in mri data. IEEE transactions on medical imaging , 17(1):87–97

    1998

  30. [38]

    M., Jenkinson, M., Woolrich, M

    Smith, S. M., Jenkinson, M., Woolrich, M. W., Beckmann, C. F., Behrens, T. E., Johansen-Berg, H., Bannister, P. R., De Luca, M., Drob- njak, I., Flitney, D. E., Niazy, R. K., Saunders, J., Vickers, J., Zhang, Y., De Stefano, N., Brady, J. M., and Matthews, P. M. (2004). Advanc...

    2004

  31. [39]

    Takao, H., Hayashi, N., and Ohtomo, K. (2011). Effect of scanner in lon- gitudinal studies of brain volume changes. Journal of Magnetic Resonance Imaging, 34(2):438–444

    2011

  32. [40]

    Tanaka, S. C., Yamashita, A., Yahata, N., Itahashi, T., Lisi, G., Ya- mada, T., Ichikawa, N., Takamura, M., Yoshihara, Y., Kunimatsu, A., Okada, N., Hashimoto, R., Okada, G., Sakai, Y., Morimoto, J., Naru- moto, J., Shimada, Y., Mano, H., Yoshida, W., Seymour, B., Shimizu, T.,...

    2021

  33. [41]

    E., Minhas, D

    Torbati, M. E., Minhas, D. S., Ahmad, G., O’Connor, E. E., Muschelli, J., Laymon, C. M., Yang, Z., Cohen, A. D., Aizenstein, H. J., Klunk, W. E., Christian, B. T., Hwang, S. J., Crainiceanu, C. M., and Tudorascu, D. L. (2021). A multi-scanner neuroimaging data harmonization us...

    2021

  34. [42]

    J., Avants, B

    Tustison, N. J., Avants, B. B., Cook, P. A., Zheng, Y., Egan, A., Yushke- vich, P. A., and Gee, J. C. (2010). N4itk: improved n3 bias correction. IEEE transactions on medical imaging , 29(6):1310–1320. 62

    2010

  35. [43]

    Villani, C. (2008). Optimal transport: old and new , volume 338. Springer

    2008

  36. [44]

    B., Fjell, A

    Walhovd, K. B., Fjell, A. M., Reinvang, I., Lundervold, A., Dale, A. M., Eilertsen, D. E., Quinn, B. T., Salat, D., Makris, N., and Fischl, B. (2005). Effects of age on volumes of cortex, white matter and subcortical struc- tures. Neurobiology of aging, 26(9):1261–1270

    2005

  37. [45]

    Wang, P., Zheng, W., Chen, T., and Wang, Z. (2022). Anti- oversmoothing in deep vision transformers via the fourier domain analysis: From theory to practice. arXiv preprint arXiv:2203.05962

    2022 arXiv

  38. [46]

    C., Sheikh, H

    Wang, Z., Bovik, A. C., Sheikh, H. R., and Simoncelli, E. P. (2004). Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing , 13(4):600–612

    2004

  39. [47]

    Wei, D., Zhuang, K., Chen, Q., Yang, W., Liu, W., Wang, K., Sun, J., and Qiu, J. (2017). Structural and functional mri from a cross-sectional southwest university adult lifespan dataset (sald). bioRxiv, page 177279

    2017

  40. [48]

    A., Shechtman, E., and Wang, O

    Zhang, R., Isola, P., Efros, A. A., Shechtman, E., and Wang, O. (2018). The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 586–595

    2018

  41. [49]

    E., Liu, Y., He, Y., Newsome, S

    Zuo, L., Dewey, B. E., Liu, Y., He, Y., Newsome, S. D., Mowry, E. M., Resnick, S. M., Prince, J. L., and Carass, A. (2021). Unsupervised mr harmonization by learning disentangled representations using information bottleneck theory. NeuroImage, 243:118569

    2021

  42. [50]

    Zuo, X.-N., Xu, T., and Milham, M. P. (2019). Harnessing reliability for neuroscience research. Nature human behaviour, 3(8):768–771. 63

    2019

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.