pith. machine review for the scientific record. sign in

arxiv: 2604.18781 · v1 · submitted 2026-04-20 · 💻 cs.CV

Recognition: unknown

CAHAL: Clinically Applicable resolution enHAncement for Low-resolution MRI scans

Sergio Morell-Ortega , \'Angela Gonz\'alez-Cebri\'an , Boris Mansencal , Marien Gadea , Roberto Vivo-Hernando , Gregorio Rubio , Fernando Aparici , Maria de la Iglesia-Vaya
show 3 more authors
Gwenaelle Catheline Pierrick Coup\'e Jos\'e V. Manj\'on
Authors on Pith no claims yet

Pith reviewed 2026-05-10 05:22 UTC · model grok-4.3

classification 💻 cs.CV
keywords MRI super-resolutionhallucination-robustphysics-informedmixture of expertsclinical brain imagingmorphometric analysisT1-weighted FLAIR
0
0 comments X

The pith

CAHAL enhances low-resolution clinical MRI scans without introducing anatomical hallucinations or volumetric distortions.

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

The paper proposes CAHAL as a way to turn routine thick-slice, anisotropic brain MRI into higher-resolution volumes that remain faithful to the original anatomy. Existing generative super-resolution approaches often produce visually appealing results but create false structures and measurement errors that break downstream quantitative analysis. CAHAL instead stays in the patient's native acquisition space and routes each scan through experts that are conditioned on the actual resolution and slice spacing seen in real clinical data. Training uses physics-based degradations drawn from large real-world databases rather than synthetic pairs. The result is a method that the authors show improves accuracy and speed on T1-weighted and FLAIR sequences while reducing the risks that make prior techniques unsafe for clinical morphometry.

Core claim

CAHAL is a hallucination-robust, physics-informed resolution enhancement framework that operates directly in the patient's native acquisition space. It employs a deterministic bivariate Mixture of Experts architecture with residual 3D U-Net experts conditioned on both volumetric resolution and acquisition anisotropy. These experts are optimized with a composite loss that combines edge-penalised spatial reconstruction, Fourier-domain spectral coherence matching, and segmentation-guided semantic consistency. Training pairs are generated on-the-fly via physics-based degradation sampled from a large-scale real-world database. When tested on T1-weighted and FLAIR sequences, CAHAL outperforms the

What carries the argument

Deterministic bivariate Mixture of Experts routing each input through specialised residual 3D U-Net experts conditioned on resolution and anisotropy, trained with edge-penalised reconstruction, Fourier spectral coherence, and segmentation-guided consistency.

If this is right

  • Routine thick-slice clinical MRI acquisitions can support accurate automated morphometric analysis without requiring new isotropic protocols.
  • Super-resolved volumes avoid the systematic overestimation and structural distortions that compromise diagnostic safety in generative methods.
  • The on-the-fly training strategy enables robust performance across varying clinical scanners and protocols not seen during development.
  • Efficiency improvements make large-scale processing of existing hospital archives feasible for quantitative research.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The conditioning on both resolution and anisotropy could be adapted to other anisotropic modalities such as CT or ultrasound to reduce similar hallucination risks.
  • Embedding the method in clinical PACS systems might allow real-time quality checks that flag when enhancement reliability drops for a given acquisition.
  • Longitudinal studies using CAHAL on legacy low-resolution scans could reveal new biomarkers that were previously inaccessible due to resolution limits.

Load-bearing premise

That on-the-fly physics-based degradation pairs drawn from a real-world database will produce models that generalize without hallucinations or distortions to all unseen clinical acquisitions.

What would settle it

A side-by-side comparison of CAHAL outputs against true high-resolution follow-up scans of the same patients, measuring whether volumetric estimates match and whether any anatomically implausible structures appear.

Figures

Figures reproduced from arXiv: 2604.18781 by \'Angela Gonz\'alez-Cebri\'an, Boris Mansencal, Fernando Aparici, Gregorio Rubio, Gwenaelle Catheline, Jos\'e V. Manj\'on, Maria de la Iglesia-Vaya, Marien Gadea, Pierrick Coup\'e, Roberto Vivo-Hernando, Sergio Morell-Ortega.

Figure 1
Figure 1. Figure 1: Left/right areas are some examples of real case resolutions and appearance. In the centre, the distribution of [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Progressive transfer learning strategy across the bivariate expert manifold. The matrix illustrates the training [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Visual comparison between the segmentations predicted by our DPN (left) and the proxy ground truth labels [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: PSNR trend across the two-dimensional acquisition manifold. Anisotropy is mapped into three discrete [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative assessment of cerebellar reconstruction of a case from volumetric cluster 5 anisotropy 1 (1x1x5 [PITH_FULL_IMAGE:figures/full_fig_p014_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Example reconstruction case of 1x1x7 mm3 with SuperSynth and CAHAL. To assess practical deployability, we evaluated CAHAL’s computational efficiency across a range of hardware con￾figurations. On a 32 GB VRAM GPU, CAHAL processes a standard 200 × 200 × 200 volume in approximately 3.5 seconds, compared to 30 seconds for SuperSynth under identical conditions. To support deployment on standard hardware, CAHAL… view at source ↗
Figure 7
Figure 7. Figure 7: Impact of resolution improvement on automated registration. The proposed method was found to be [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Preservation of structural segmentation accuracy. Subplots display the Dice for: whole-brain segmentation [PITH_FULL_IMAGE:figures/full_fig_p018_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Generative volumetric bias. Statistical significance between method pairs within each diagnosis group (CN or [PITH_FULL_IMAGE:figures/full_fig_p018_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Preservation of diagnostic power and signal fidelity. [PITH_FULL_IMAGE:figures/full_fig_p019_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Validation on ultra-high resolution cerebellar structures. Subplots display the Dice for: whole-brain [PITH_FULL_IMAGE:figures/full_fig_p020_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Validation on white matter lesion segmentation. Left: average Dice of 132 brain structures. Right: average [PITH_FULL_IMAGE:figures/full_fig_p021_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Demographic profile of the curated training cohorts. Histograms illustrate the age distribution hued by biological sex. The combined datasets provide comprehensive coverage across the human lifespan, ensuring the model’s exposure to age-related structural variance and diverse brain morphologies [PITH_FULL_IMAGE:figures/full_fig_p027_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Qualitative evaluation of a moderately anisotropic clinical acquisition ( [PITH_FULL_IMAGE:figures/full_fig_p028_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Impact of resolution enhancement on automated disease grading spatial maps (DeepGrading score) Nguyen [PITH_FULL_IMAGE:figures/full_fig_p029_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Bland-Altman analysis of diagnostic signal agreement (grading score). Comparison of native LR, our [PITH_FULL_IMAGE:figures/full_fig_p030_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Qualitative reconstruction of severe through-plane anisotropic degradation ( [PITH_FULL_IMAGE:figures/full_fig_p030_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Preservation of pathological segmentation boundaries of a [PITH_FULL_IMAGE:figures/full_fig_p031_18.png] view at source ↗
read the original abstract

Large-scale automated morphometric analysis of brain MRI is limited by the thick-slice, anisotropic acquisitions prevalent in routine clinical practice. Existing generative super-resolution (SR) methods produce visually compelling isotropic volumes but often introduce anatomical hallucinations, systematic volumetric overestimation, and structural distortions that compromise downstream quantitative analysis and diagnostic safety. To address this, we propose CAHAL (Clinically Applicable resolution enHAncement for Low-resolution MRI scans), a hallucination-robust, physics-informed resolution enhancement framework that operates directly in the patient's native acquisition space. CAHAL employs a deterministic bivariate Mixture of Experts (MoE) architecture routing each input through specialised residual 3D U-Net experts conditioned on both volumetric resolution and acquisition anisotropy, two independent descriptors of clinical MRI acquisition. Experts are optimised with a composite loss combining edge-penalised spatial reconstruction, Fourier-domain spectral coherence matching, and a segmentation-guided semantic consistency constraint. Training pairs are generated on-the-fly via physics-based degradation sampled from a large-scale real-world database, ensuring robust generalisation. Validated on T1-weighted and FLAIR sequences against generative baselines, CAHAL achieves state-of-the-art results, improving the best related methods in terms of accuracy and efficiency.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The manuscript introduces CAHAL, a hallucination-robust super-resolution method for low-resolution clinical brain MRI. It employs a deterministic bivariate Mixture of Experts architecture with residual 3D U-Net experts conditioned on volumetric resolution and acquisition anisotropy. Experts are trained via a composite loss (edge-penalised spatial reconstruction, Fourier spectral coherence, and segmentation-guided semantic consistency) on pairs generated on-the-fly through physics-based degradation sampled from real-world data. The paper claims state-of-the-art performance on T1-weighted and FLAIR sequences relative to generative baselines, with gains in accuracy and efficiency.

Significance. If the quantitative results and hallucination assessments hold, the work addresses a clinically important gap by enabling reliable isotropic volumes from routine thick-slice acquisitions without introducing volumetric biases or structural distortions that affect downstream morphometry. The physics-informed on-the-fly degradation sampling and multi-fidelity composite loss are explicit strengths that support generalisation claims.

major comments (1)
  1. Abstract: the central claim that CAHAL 'achieves state-of-the-art results, improving the best related methods in terms of accuracy and efficiency' and is 'hallucination-robust' is not accompanied by any quantitative metrics (e.g., PSNR, SSIM, Dice, volumetric error), error bars, statistical tests, or details on hallucination measurement. This evidence is load-bearing for the primary contribution and cannot be evaluated from the provided text.
minor comments (1)
  1. The phrase 'deterministic bivariate Mixture of Experts (MoE) routing conditioned on resolution and anisotropy' would benefit from an explicit description of the routing function and any conditioning mechanism in the methods section to confirm it introduces no additional free parameters.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback and positive evaluation of the clinical significance of our work. We address the major comment below.

read point-by-point responses

  1. Referee: Abstract: the central claim that CAHAL 'achieves state-of-the-art results, improving the best related methods in terms of accuracy and efficiency' and is 'hallucination-robust' is not accompanied by any quantitative metrics (e.g., PSNR, SSIM, Dice, volumetric error), error bars, statistical tests, or details on hallucination measurement. This evidence is load-bearing for the primary contribution and cannot be evaluated from the provided text.

    Authors: We agree that the abstract should include key quantitative evidence to support the central claims. The full manuscript reports detailed results in the Experiments section, including PSNR and SSIM for reconstruction accuracy, Dice scores and volumetric error for downstream segmentation and morphometry, error bars across test cases, and statistical comparisons against baselines. Hallucination robustness is quantified via the segmentation-guided loss, Fourier coherence, and absence of systematic volumetric biases. To make the abstract self-contained, we will revise it to incorporate representative metric values (e.g., relative improvements in PSNR/SSIM and Dice) and a brief note on the hallucination assessment approach. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper introduces a new MoE-based architecture and on-the-fly physics-informed training procedure without equations, derivations, or predictions that reduce by construction to fitted parameters or self-citations. All load-bearing elements (expert routing, composite loss, degradation sampling) are externally defined and validated against independent baselines on T1/FLAIR data, rendering the derivation self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

Only the abstract is available, so the ledger is necessarily incomplete. The central claim rests on standard U-Net residual blocks, mixture-of-experts routing, and physics-based degradation simulation whose details are not visible.

axioms (2)
  • domain assumption Residual 3D U-Net blocks form suitable expert networks for volumetric MRI super-resolution
    Invoked by the choice of expert architecture in the abstract.
  • domain assumption Fourier-domain spectral coherence matching preserves anatomical fidelity
    Part of the composite loss described in the abstract.
invented entities (1)
  • Deterministic bivariate Mixture of Experts (MoE) routing conditioned on resolution and anisotropy no independent evidence
    purpose: Route each low-resolution input to specialised experts based on two clinical acquisition descriptors
    Core architectural innovation stated in the abstract; no independent evidence provided beyond the claim itself.

pith-pipeline@v0.9.0 · 5571 in / 1403 out tokens · 37625 ms · 2026-05-10T05:22:41.210559+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

294 extracted references · 237 canonical work pages

  1. [1]

    2018 , journal =

    Dolz, Jose and Desrosiers, Christian and Ben Ayed, Ismail , month =. 2018 , journal =. doi:10.1016/J.NEUROIMAGE.2017.04.039 , issn =

    work page doi:10.1016/j.neuroimage.2017.04.039 2018

  2. [2]

    Wong, Ken C. L. and Moradi, Mehdi and Tang, Hui and Syeda-Mahmood, Tanveer , month =. 2018 , journal =. doi:10.1007/978-3-030-00931-1\_70 , arxivId =

    work page doi:10.1007/978-3-030-00931-1 2018

  3. [3]

    doi:10.1007/978-3-319-46723-8\_49 , issn =

    2016 , journal =. doi:10.1007/978-3-319-46723-8\_49 , issn =

    work page doi:10.1007/978-3-319-46723-8 2016

  4. [4]

    and Cutting, Laurie E

    Huo, Yuankai and Xu, Zhoubing and Xiong, Yunxi and Aboud, Katherine and Parvathaneni, Prasanna and Bao, Shunxing and Bermudez, Camilo and Resnick, Susan M. and Cutting, Laurie E. and Landman, Bennett A. , month =. 2019 , journal =. doi:10.1016/J.NEUROIMAGE.2019.03.041 , issn =

    work page doi:10.1016/j.neuroimage.2019.03.041 2019

  5. [5]

    2019 , journal =

    Liu, Xiaoxuan and Faes, Livia and Kale, Aditya U and Wagner, Siegfried K and Fu, Dun Jack and Bruynseels, Alice and Mahendiran, Thushika and Moraes, Gabriella and Shamdas, Mohith and Kern, Christoph and Ledsam, Joseph R and Schmid, Martin K and Balaskas, Konstantinos and Topol, Eric J and Bachmann, Lucas M and Keane, Pearse A and Denniston, Alastair K , n...

    work page doi:10.1016/s2589-7500(19)30123-2 2019

  6. [6]

    and Price, Anthony N

    Schlemper, Jo and Caballero, Jose and Hajnal, Joseph V. and Price, Anthony N. and Rueckert, Daniel , number =. 2017 , journal =. doi:10.1109/TMI.2017.2760978 , issn =

    work page doi:10.1109/tmi.2017.2760978 2017

  7. [7]

    Monteagudo, Carlos , arxivId =

  8. [8]

    2022 , journal =

    Martinini, Filippo and Mangia, Mauro and Marchioni, Alex and Rovatti, Riccardo and Setti, Gianluca , number =. 2022 , journal =. doi:10.1109/JSTSP.2022.3171082 , issn =

    work page doi:10.1109/jstsp.2022.3171082 2022

  9. [9]

    , number =

    Yeung, Daniel and McKenzie, Craig and Indelicato, Daniel J. , number =. 2014 , journal =. doi:10.1002/PBC.24593 , issn =

    work page doi:10.1002/pbc.24593 2014

  10. [10]

    2021 , journal =

    Al-Rajab, Murad and Lu, Joan and Xu, Qiang , number =. 2021 , journal =. doi:10.1371/JOURNAL.PONE.0249094 , issn =

    work page doi:10.1371/journal.pone.0249094 2021

  11. [11]

    2022 , journal =

    Yousef, Rammah and Gupta, Gaurav and Yousef, Nabhan and Khari, Manju , number =. 2022 , journal =. doi:10.1007/S00530-021-00884-5 , issn =

    work page doi:10.1007/s00530-021-00884-5 2022

  12. [12]

    Puckett, W. R. and Lightfoot, E. N. , number =. 1995 , journal =. doi:10.1152/AJPENDO.1995.269.6.E1115 , issn =

    work page doi:10.1152/ajpendo.1995.269.6.e1115 1995

  13. [14]

    Lehmann, E. D. and Deutsch, T. , number =. 1992 , journal =. doi:10.1016/0141-5425(92)90058-S , issn =

    work page doi:10.1016/0141-5425(92)90058-s 1992

  14. [15]

    and Flavell, Jonathan and Cussans, Emma and Ramnani, Narender , number =

    Diedrichsen, Jörn and Balsters, Joshua H. and Flavell, Jonathan and Cussans, Emma and Ramnani, Narender , number =. 2009 , journal =. doi:10.1016/J.NEUROIMAGE.2009.01.045 , issn =

    work page doi:10.1016/j.neuroimage.2009.01.045 2009

  15. [16]

    and Afraites, L

    Nachaoui, M. and Afraites, L. and Laghrib, A. , month =. 2021 , journal =. doi:10.1016/J.IMAGE.2021.116505 , issn =

    work page doi:10.1016/j.image.2021.116505 2021

  16. [18]

    2022 , journal =

    Mohan, Radhe and Radhe Mohan, Correspondence , number =. 2022 , journal =. doi:10.1002/PRO6.1149 , issn =

    work page doi:10.1002/pro6.1149 2022

  17. [19]

    and Nyholm, Tufve , number =

    Edmund, Jens M. and Nyholm, Tufve , number =. 2017 , journal =. doi:10.1186/S13014-016-0747-Y/FIGURES/6 , issn =

    work page doi:10.1186/s13014-016-0747-y/figures/6 2017

  18. [20]

    2006 , journal =

    Diedrichsen, Jörn , number =. 2006 , journal =. doi:10.1016/J.NEUROIMAGE.2006.05.056 , issn =

    work page doi:10.1016/j.neuroimage.2006.05.056 2006

  19. [21]

    2018 , journal =

    Min, Erxue and Guo, Xifeng and Liu, Qiang and Zhang, Gen and Cui, Jianjing and Long, Jun , month =. 2018 , journal =. doi:10.1109/ACCESS.2018.2855437 , keywords =

    work page doi:10.1109/access.2018.2855437 2018

  20. [22]

    and Yong Park, Sung and Klein, Eric and Flanz, Jacob and Yorke, Ellen D

    Arjomandy, Bijan and Taylor, Paige and Ainsley, Christopher and Safai, Sairos and Sahoo, Narayan and Pankuch, Mark and Farr, Jonathan B. and Yong Park, Sung and Klein, Eric and Flanz, Jacob and Yorke, Ellen D. and Followill, David and Kase, Yuki , number =. 2019 , journal =. doi:10.1002/MP.13622 , issn =

    work page doi:10.1002/mp.13622 2019

  21. [23]

    2008 , journal =

    R, Rajaganeshan and CL, Ludlam and DP, Francis and SV, Parasramka and R, Sutton , number =. 2008 , journal =. doi:10.1111/J.1742-1241.2007.01390..X , issn =

    work page doi:10.1111/j.1742-1241.2007.01390..x 2008

  22. [24]

    and Ludlam, C

    Rajaganeshan, R. and Ludlam, C. L. and Francis, D. P. and Parasramka, S. V. and Sutton, R. , number =. 2007 , journal =

    2007

  23. [25]

    2017 , journal =

    Fornes-Leal, Alejandro and Garcia-Pardo, Concepcion and Cardona, Narcís and Castello-Palacios, Sergio and Valles-Lluch, Ana , month =. 2017 , journal =. doi:10.1109/ISMICT.2017.7891761 , keywords =

    work page doi:10.1109/ismict.2017.7891761 2017

  24. [26]

    ieeexplore.ieee.org , author =

  25. [27]

    , number =

    Chen, Peng Jen and Lin, Meng Chiung and Lai, Mei Ju and Lin, Jung Chun and Lu, Henry Horng Shing and Tseng, Vincent S. , number =. 2018 , journal =. doi:10.1053/J.GASTRO.2017.10.010 , issn =

    work page doi:10.1053/j.gastro.2017.10.010 2018

  26. [28]

    and Ba, Jimmy Lei , month =

    Kingma, Diederik P. and Ba, Jimmy Lei , month =. 2014 , journal =

    2014

  27. [29]

    Adaptivenon-localmeansdenoisingofmrimageswithspatially varying noise levels

    Manj. 2010 , journal =. doi:10.1002/JMRI.22003 , issn =

    work page doi:10.1002/jmri.22003 2010

  28. [30]

    and Lege, David and Mohan, Radhe and Komaki, Ritsuko and Cox, James D

    Koay, Eugene J. and Lege, David and Mohan, Radhe and Komaki, Ritsuko and Cox, James D. and Chang, Joe Y. , number =. 2012 , journal =. doi:10.1016/J.IJROBP.2012.02.041 , issn =

    work page doi:10.1016/j.ijrobp.2012.02.041 2012

  29. [31]

    and Song, Gang , month =

    Avants, Brian and Tustison, Nicholas J. and Song, Gang , month =. 2009 , journal =

    2009

  30. [32]

    and Kercher, Matthew J

    Kulubya, Edwin S. and Kercher, Matthew J. and Phillips, H. Westley and Antony, Reuben and Edwards, Michael S.B. , number =. 2022 , journal =. doi:10.3390/CHILDREN10010062 , issn =

    work page doi:10.3390/children10010062 2022

  31. [33]

    2002 , journal =

    Allen, J and Murray, A , number =. 2002 , journal =. doi:10.1038/sj.jhh.1001478 , issn =

    work page doi:10.1038/sj.jhh.1001478 2002

  32. [34]

    and Etoz, Sevde , number =

    Brace, Christopher L. and Etoz, Sevde , number =. 2020 , journal =. doi:10.3390/s20185372 , issn =

    work page doi:10.3390/s20185372 2020

  33. [35]

    and Federspil, G

    Cobelli, C. and Federspil, G. and Pacini, G. and Salvan, A. and Scandellari, C. , number =. 1982 , journal =. doi:10.1016/0025-5564(82)90050-5 , issn =

    work page doi:10.1016/0025-5564(82)90050-5 1982

  34. [36]

    2015 , journal =

    O'Shea, Keiron and Nash, Ryan , number =. 2015 , journal =. doi:10.22214/ijraset.2022.47789 , arxivId =

    work page doi:10.22214/ijraset.2022.47789 2015

  35. [37]

    2016 , journal =

    Giraud, Rémi and Ta, Vinh Thong and Papadakis, Nicolas and Manj. 2016 , journal =. doi:10.1016/J.NEUROIMAGE.2015.07.076 , issn =

    work page doi:10.1016/j.neuroimage.2015.07.076 2016

  36. [38]

    Ruder, Sebastian , url =

  37. [39]

    2018 , journal =

    T, Hirasawa and K, Aoyama and T, Tanimoto and S, Ishihara and S, Shichijo and T, Ozawa and T, Ohnishi and M, Fujishiro and K, Matsuo and J, Fujisaki and T, Tada , number =. 2018 , journal =. doi:10.1007/S10120-018-0793-2 , issn =

    work page doi:10.1007/s10120-018-0793-2 2018

  38. [40]

    and Jedynak, Bruno and Rigg, Rachel and Du, Annie and Landman, Bennett A

    Bogovic, John A. and Jedynak, Bruno and Rigg, Rachel and Du, Annie and Landman, Bennett A. and Prince, Jerry L. and Ying, Sarah H. , number =. 2013 , journal =. doi:10.1016/J.NEUROIMAGE.2012.08.075 , issn =

    work page doi:10.1016/j.neuroimage.2012.08.075 2013

  39. [41]

    2021 , journal =

    Tang, Yunbo and Chen, Dan and Li, Xiaoli , number =. 2021 , journal =. doi:10.1145/3448302 , keywords =

    work page doi:10.1145/3448302 2021

  40. [42]

    Barragán-Montero, U

    Barrag. 2021 , journal =. doi:10.1016/J.EJMP.2021.04.016 , issn =

    work page doi:10.1016/j.ejmp.2021.04.016 2021

  41. [43]

    , month =

    Singh, Ramandeep and Wu, Weiwen and Wang, Ge and Kalra, Mannudeep K. , month =. 2020 , journal =. doi:10.1016/J.EJMP.2020.11.012 , issn =

    work page doi:10.1016/j.ejmp.2020.11.012 2020

  42. [44]

    2020 , journal =

    Coup. 2020 , journal =. doi:10.1016/J.NEUROIMAGE.2020.117026 , issn =

    work page doi:10.1016/j.neuroimage.2020.117026 2020

  43. [45]

    doi:10.1007/s00259-015-3137-z , keywords =

    Brendle, C and Schwenzer, N F and Rempp, & H and Schmidt, & H and Pfannenberg, & C and La Foug. doi:10.1007/s00259-015-3137-z , keywords =

    work page doi:10.1007/s00259-015-3137-z

  44. [46]

    2022 , journal =

    Sinclair, Matthew and Schuh, Andreas and Hahn, Karl and Petersen, Kersten and Bai, Ying and Batten, James and Schaap, Michiel and Glocker, Ben , month =. 2022 , journal =. doi:10.1016/J.MEDIA.2022.102383 , issn =

    work page doi:10.1016/j.media.2022.102383 2022

  45. [47]

    and Welling, Max , month =

    Kingma, Diederik P. and Welling, Max , month =. 2013 , journal =

    2013

  46. [48]

    and Carass, Aaron and Jedynak, Bruno M

    Yang, Zhen and Ye, Chuyang and Bogovic, John A. and Carass, Aaron and Jedynak, Bruno M. and Ying, Sarah H. and Prince, Jerry L. , month =. 2016 , journal =. doi:10.1016/J.NEUROIMAGE.2015.09.032 , issn =

    work page doi:10.1016/j.neuroimage.2015.09.032 2016

  47. [49]

    and Shukla, K

    Sharma, Neeraj and Ray, Amit K. and Shukla, K. K. and Sharma, Shiru and Pradhan, Satyajit and Srivastva, Arvind and Aggarwal, Lalit , number =. 2010 , journal =. doi:10.4103/0971-6203.58777 , issn =

    work page doi:10.4103/0971-6203.58777 2010

  48. [50]

    , month =

    Han, Shuo and Carass, Aaron and He, Yufan and Prince, Jerry L. , month =. 2020 , journal =. doi:10.1016/J.NEUROIMAGE.2020.116819 , issn =

    work page doi:10.1016/j.neuroimage.2020.116819 2020

  49. [51]

    and Shukla, Amit K

    Ezugwu, Absalom E. and Shukla, Amit K. and Agbaje, Moyinoluwa B. and Oyelade, Olaide N. and Jos. 2021 , journal =. doi:10.1007/S00521-020-05395-4 , keywords =

    work page doi:10.1007/s00521-020-05395-4 2021

  50. [52]

    2017 , journal =

    R, Zhang and Y, Zheng and TW, Mak and R, Yu and SH, Wong and JY, Lau and CC, Poon , number =. 2017 , journal =. doi:10.1109/JBHI.2016.2635662 , issn =

    work page doi:10.1109/jbhi.2016.2635662 2017

  51. [53]

    2021 , journal =

    Wang, Tao and Lu, Changhua and Sun, Yining and Yang, Mei and Liu, Chun and Ou, Chunsheng , number =. 2021 , journal =. doi:10.3390/e23010119 , issn =

    work page doi:10.3390/e23010119 2021

  52. [54]

    2015 , journal =

    Ioffe, Sergey and Szegedy, Christian , month =. 2015 , journal =

    2015

  53. [55]

    2016 , journal =

    Luo, Jake and Wu, Min and Gopukumar, Deepika and Zhao, Yiqing , month =. 2016 , journal =. doi:10.4137/BII.S31559 , issn =

    work page doi:10.4137/bii.s31559 2016

  54. [56]

    and Al-Jumaily, A

    Lan, H. and Al-Jumaily, A. M. and Hing, W. and Lowe, A. , pages =. 2010 , journal =

    2010

  55. [57]

    Kahn, Jr

    Smith, Barry and Arabandi, Sivaram and Brochhausen, Mathias and Calhoun, Michael and Ciccarese, Paolo and Doyle, Scott and Gibaud, Bernard and Goldberg, Ilya and Charles E. Kahn, Jr. and Overton, James and Tomaszewski, John and Gurcan, Metin , number =. 2015 , journal =. doi:10.4103/2153-3539.159214 , issn =

    work page doi:10.4103/2153-3539.159214 2015

  56. [58]

    Hupfeld, K. E. and Mcgregor, H. R. and Koppelmans, V. and Beltran, N. E. and Kofman, I. S. and De Dios, Y. E. and Riascos, R. F. and Reuter-Lorenz, P. A. and Wood, S. J. and Bloomberg, J. J. and Mulavara, A. P. and Seidler, R. D. , number =. 2022 , journal =. doi:10.1093/CERCOR/BHAB239 , issn =

    work page doi:10.1093/cercor/bhab239 2022

  57. [59]

    Sobhaninia, Zahra and Karimi, Nader and Khadivi, Pejman and Samavi, Shadrokh , isbn =

  58. [60]

    and Adedigba, Adeyinka P

    Adeshina, Steve A. and Adedigba, Adeyinka P. and Adeniyi, Ahmed A. and Aibinu, Abiodun M. , month =. 2019 , journal =. doi:10.1109/ICECCO.2018.8634690 , keywords =

    work page doi:10.1109/icecco.2018.8634690 2019

  59. [61]

    Sharma, S. N. , number =. 1991 , journal =. doi:10.1056/NEJMVCM1313338 , issn =

    work page doi:10.1056/nejmvcm1313338 1991

  60. [62]

    Singanamalli, Asha and Wang, Haibo and Madabhushi, Anant and Weiner, Michael and Aisen, Paul and Petersen, Ronald and Jack, Clifford and Jagust, William and Trojanowki, John and Toga, Arthur and Beckett, Laurel and Green, Robert and Saykin, Andrew and Morris, John and Shaw, Leslie and Kaye, Jeffrey and Quinn, Joseph and Silbert, Lisa and Lind, Betty and C...

    work page doi:10.1038/s41598-017-03925-0 2017

  61. [63]

    2016 , journal =

    Cobe. 2016 , journal =. doi:10.1016/J.RARD.2015.09.006 , issn =

    work page doi:10.1016/j.rard.2015.09.006 2016

  62. [64]

    Laidi, Charles and Floris, Dorothea L. and Tillmann, Julian and Elandaloussi, Yannis and Zabihi, Mariam and Charman, Tony and Wolfers, Thomas and Durston, Sarah and Moessnang, Carolin and Dell'Acqua, Flavio and Ecker, Christine and Loth, Eva and Murphy, Declan and Baron-Cohen, Simon and Buitelaar, Jan K. and Marquand, Andre F. and Beckmann, Christian F. a...

    work page doi:10.1016/j.biopsych.2022.05.020 2022

  63. [65]

    and Hajek, T

    Laidi, C. and Hajek, T. and Spaniel, F. and Kolenic, M. and d'Albis, M. A. and Sarrazin, S. and Mangin, J. F. and Duchesnay, E. and Brambilla, P. and Wessa, M. and Linke, J. and Polosan, M. and Favre, P. and Versace, A. L. and Phillips, M. L. and Manjon, J. V. and Romero, J. E. and Hozer, F. and Leboyer, M. and Coupe, P. and Houenou, J. , number =. 2019 ,...

    work page doi:10.1111/acps.13087 2019

  64. [66]

    and d'Albis, M

    Laidi, C. and d'Albis, M. A. and Wessa, M. and Linke, J. and Phillips, M. L. and Delavest, M. and Bellivier, F. and Versace, A. and Almeida, J. and Sarrazin, S. and Poupon, C. and Le Dudal, K. and Daban, C. and Hamdani, N. and Leboyer, M. and Houenou, J. , number =. 2015 , journal =. doi:10.1111/ACPS.12363 , issn =

    work page doi:10.1111/acps.12363 2015

  65. [67]

    2022 , journal =

    Faber, Jennifer and K. 2022 , journal =. doi:10.1016/J.NEUROIMAGE.2022.119703 , issn =

    work page doi:10.1016/j.neuroimage.2022.119703 2022

  66. [68]

    and Coup

    Romero, Jose E. and Coup. 2017 , journal =. doi:10.1016/J.NEUROIMAGE.2016.11.003 , issn =

    work page doi:10.1016/j.neuroimage.2016.11.003 2017

  67. [69]

    2023 , journal =

    Salazar, Ana Paula and Mcgregor, Heather R and Hupfeld, Kathleen E and Beltran, Nichole E and Kofman, Igor S and De Dios, Yiri E and Riascos, Roy F and Reuter-Lorenz, Patricia A and Bloomberg, Jacob J and Mulavara, Ajitkumar P and Wood, Scott J and Seidler, Rachael D , number =. 2023 , journal =. doi:10.1093/CERCOR/BHAC232 , issn =

    work page doi:10.1093/cercor/bhac232 2023

  68. [70]

    2017 , booktitle =

    Srinivas, Suraj and Sarvadevabhatla, Ravi K and Mopuri, Konda R and Prabhu, Nikita and Kruthiventi, Srinivas S S and Babu, R Venkatesh , pages =. 2017 , booktitle =

    2017

  69. [71]

    and Taake, S

    Bayerl, C. and Taake, S. and Moll, I. and Jung, E. G. , number =. 1995 , journal =. doi:10.1111/J.1600-0781.1995.TB00157.X , issn =

    work page doi:10.1111/j.1600-0781.1995.tb00157.x 1995

  70. [72]

    Geddes, L. A. and Voelz, M. and Combs, C. and Reiner, D. and Babbs, C. F. , number =. 1982 , journal =. doi:10.1007/BF02367308 , issn =

    work page doi:10.1007/bf02367308 1982

  71. [73]

    and Wang, Zhen and Wong, William W

    Patel, Samir H. and Wang, Zhen and Wong, William W. and Murad, Mohammad Hassan and Buckey, Courtney R. and Mohammed, Khaled and Alahdab, Fares and Altayar, Osama and Nabhan, Mohammed and Schild, Steven E. and Foote, Robert L. , number =. 2014 , journal =. doi:10.1016/S1470-2045(14)70268-2 , issn =

    work page doi:10.1016/s1470-2045(14)70268-2 2014

  72. [74]

    2000 , journal =

    Ito, Shosuke and Wakamatsu, Kazumasa and Ozeki, Hiroyuki , number =. 2000 , journal =. doi:10.1034/J.1600-0749.13.S8.19.X , issn =

    work page doi:10.1034/j.1600-0749.13.s8.19.x 2000

  73. [75]

    2002 , journal =

    Stewart, Lesley and Burdett, Sarah and. 2002 , journal =. doi:10.1002/14651858.CD003913/MEDIA/CDSR/CD003913/IMAGE\_N/NCD003913-CMP-001-07.PNG , issn =

    work page doi:10.1002/14651858.cd003913/media/cdsr/cd003913/image 2002

  74. [76]

    2017 , journal =

    Nies, Hui Wen and Daud, Kauthar Mohd and Remli, Muhammad Akmal and Mohamad, Mohd Saberi and Deris, Safaai and Omatu, Sigeru and Kasim, Shahreen and Sulong, Ghazali , pages =. 2017 , journal =. doi:10.1007/978-3-319-60816-7\_8 , keywords =

    work page doi:10.1007/978-3-319-60816-7 2017

  75. [77]

    and Ahuja, Chirag K

    Singla, Aditya Kumar and Madan, Renu and Gupta, Kirti and Goyal, Shikha and Kumar, Narendra and Sahoo, Sushant Kumar and Uppal, Deepak K. and Ahuja, Chirag K. , number =. 2021 , journal =. doi:10.3857/ROJ.2020.00591 , issn =

    work page doi:10.3857/roj.2020.00591 2021

  76. [78]

    2020 , journal =

    Ram. 2020 , journal =. doi:10.1007/S00109-020-01874-2 , issn =

    work page doi:10.1007/s00109-020-01874-2 2020

  77. [79]

    and Deelchand, Dinesh K

    Chandrasekaran, Jayashree and Petit, Emilien and Park, Young Woo and du Montcel, Sophie Tezenas and Joers, James M. and Deelchand, Dinesh K. and Pova. 2022 , journal =. doi:10.1002/ANA.26573 , issn =

    work page doi:10.1002/ana.26573 2022

  78. [80]

    elibrary.asabe.org , author =

  79. [81]

    2013 , journal =

    Coup. 2013 , journal =. doi:10.1016/J.NEUROIMAGE.2013.06.030 , issn =

    work page doi:10.1016/j.neuroimage.2013.06.030 2013

  80. [82]

    2017 , journal =

    Issa, Iyad A and Noureddine, Malak , number =. 2017 , journal =. doi:10.3748/WJG.V23.I28.5086 , pmid =

    work page doi:10.3748/wjg.v23.i28.5086 2017

Showing first 80 references.