Pith. sign in

REVIEW 4 major objections 6 minor 132 references

Graph Neural Network Reveals the Cortical Morphology of Local Brain Aging in Normal Cognition and Alzheimer's Disease

T0 review · 4 major / 6 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read A graph neural network trained on cortical surface morphometry can estimate local brain age at every cortical vertex, and in Alzheimer's disease the largest local age gaps appear in the parahippocampal gyrus and temporal cortex.

desk verdict Useful first GNN-based vertex-level local brain-age model with credible cortical aging maps, but the abstract overclaims SOTA and MCI findings, and the AD–CN comparisons need scanner/site sensitivity checks before the biology can be trusted. read the letter →

arxiv 2601.10912 v5 pith:P4H7Q5K5 submitted 2026-01-16 q-bio.NC eess.IVq-bio.QM

Samuel D. Anderson , Jordan Jomsky , Nikhil N. Chaudhari , Nahian F. Chowdhury , Xiaoyu (Rayne) Zheng , Andrei Irimia , Alzheimers Disease Neuroimaging Initiative This is my paper
classification q-bio.NCeess.IVq-bio.QM
keywords localbrainagegapgraphneuralnetworkcorticalmorphometrythicknessAlzheimer'sdiseaseT1-weightedMRIintegratedgradients
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

This paper introduces a graph neural network that predicts a local brain-age value at every vertex of the cortical surface from five morphometric features—cortical thickness, surface area, curvature, gray/white matter intensity ratio, and sulcal depth—extracted from T1-weighted MRI. Trained on 14,250 cognitively normal adults, the model reports lower mean absolute error than existing state-of-the-art local brain-age methods on an independent Alzheimer's disease test cohort, and it identifies more biologically plausible patterns of aging. In normal aging, the largest local age gaps appear in prefrontal and parietal association cortices, consistent with the 'last-in, first-out' hypothesis of brain aging. In Alzheimer's disease, accelerated aging is widespread but strongest in the parahippocampal gyrus and surrounding temporal regions, with a region-averaged difference of about 2.7 years from controls. Regional local brain-age gaps are significantly associated with measures of Alzheimer's-related cognitive impairment, and feature attribution points to surface area and cortical thickness as the dominant contributors.

What carries the argument

The load-bearing object is a graph U-Net that turns each subject's cortical surface into a mesh graph: vertices are nodes, triangular faces are edges encoding anatomical adjacency, and five morphometric features (thickness, surface area, curvature, gray/white intensity ratio, sulcal depth) live on the nodes. Graph-convolution layers refine these features while pooling and unpooling across three standard mesh resolutions (finest ~82,000 vertices) lets information propagate across scales; the final output is one age value per vertex. A semi-global bias correction, fit on the cognitively normal test cohort, removes age-dependent systematic error before group comparisons, and integrated gradient

What would settle it

Retrain the same network on a cognitively normal sample matched to the Alzheimer's cohort for scanner, reconstruction software version, and age distribution, or estimate the bias-correction coefficients separately within the Alzheimer's cohort; if the parahippocampal gap of about 2.7 years shrinks to near zero or loses significance, the reported accelerated aging is a technical artifact.

Watch

Extended reading notes

Core claim

The central claim is that a graph U-Net operating on cortical surface meshes can estimate brain age at each of roughly 82,000 vertices from morphometric features alone. Trained on cognitively normal adults, it produces local brain-age maps in which normal aging is most pronounced in prefrontal and parietal association cortices, while Alzheimer's patients show widespread accelerated aging, with the parahippocampal gyrus differing most from controls (2.72 years region-averaged). Integrated-gradients analysis identifies surface area and cortical thickness as the main drivers; regional local age gaps correlate with cognitive decline in Alzheimer's disease.

Load-bearing premise

The load-bearing premise is that prediction errors of a model trained only on cognitively normal adults, when applied to Alzheimer's patients, reflect disease-related biological aging rather than scanner, software, age-distribution, or cohort artifacts.

Editorial extensions

If this is right

  • Local brain-age maps, rather than a single global score, let researchers see which cortical regions age faster than the rest of the brain in a given individual.
  • In Alzheimer's disease, the parahippocampal gyrus and adjacent temporal regions emerge as the clearest loci of accelerated aging, which could focus imaging biomarkers on the earliest-affected structures.
  • Regional local brain-age gaps correlate with standard measures of Alzheimer's-related cognitive impairment, so the maps carry information relevant to clinical decline.
  • Feature attribution suggests surface area and cortical thickness dominate the aging signal, guiding future feature selection and mechanistic hypotheses.

Reading between the lines

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

  • A longitudinal extension is the natural next test: repeated scans on the same Alzheimer's patients could show whether the parahippocampal local age gap grows with disease progression, which the paper does not examine.
  • If sulcal depth contributes little, a reduced feature set may reproduce the maps at lower cost; a feature-ablation experiment would settle whether the redundancy is real.
  • Because the bias correction is fit on the cognitively normal test cohort and applied to the Alzheimer's cohort, recomputing group differences with corrections estimated within each cohort would reveal how much of the 2.7-year gap is technical artifact rather than biological aging.
  • The 'last-in, first-out' pattern in normal aging suggests the same surface-native network could be applied to developmental cohorts to test whether late-maturing regions are the first to decline, an implication left implicit.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes a graph U-Net for vertex-level local brain age (LBA) estimation from cortical surface morphometry (cortical thickness, surface area, curvature, gray/white matter intensity ratio, sulcal depth). The model is trained on 14,250 cognitively normal (CN) adults from UKBB, NACC, and IXI and tested on ADNI CN (N = 1,129) and ADNI AD (N = 477). The authors report cross-validation MAE = 7.56 y, ADNI CN MAE = 7.33 y, and ADNI AD MAE = 8.15 y, with CN aging concentrated in prefrontal/parietal association cortices and AD-related accelerated aging most strongly in the parahippocampal gyrus and adjacent temporal regions. Integrated gradients identify surface area and cortical thickness as the dominant features, and regional LBA gaps are associated with several AD-relevant cognitive measures. The abstract additionally claims state-of-the-art accuracy, MCI-related aging patterns, and feature-ablation results, but these specific claims are not consistently supported by the full text.

Significance. If the central claims hold, the paper makes a useful methodological contribution: a surface-native, high-resolution LBA estimator that avoids volumetric patch assumptions and provides vertex-level interpretability through multiple morphometric features. The use of a large multi-site training sample, atlas-guided pooling, a semi-global bias-correction step, and external test data are strengths. However, the current packaging materially overstates the findings. The abstract's performance, MCI, and ablation claims conflict with the full text, and the AD–CN localization result depends on an unverified cross-dataset transfer and unadjusted scanner/site variation. These issues are load-bearing because they concern the paper's primary claims; they are addressable in revision but require additional analyses and corrected reporting.

major comments (4)
  1. [Abstract; Discussion, 'Model performance and comparison to previous work'] The abstract states that the model 'achieves lower MAE than the existing state-of-the-art' and that feature ablation highlights curvature and GWR as preferentially sensitive to AD, and the supplied abstract also reports MCI findings. The Discussion explicitly says the model 'slightly (<0.7 y) underperforms more recent variants [27]', so the state-of-the-art claim is contradicted. The full text contains no MCI cohort or MCI analysis (Table 1 lists only CN and AD; all Results and cognitive regressions use CN vs. AD), and no feature ablation is described; the interpretability analysis is integrated gradients, which Figure S3 reports as showing little AD-vs-CN difference. These unsupported claims must be corrected or removed.
  2. [Methods: Data; Results: Comparing BAGs across cohorts] The central AD–CN LBAG comparison is vulnerable to site/scanner confounding. The model is trained on UKBB/NACC/IXI and applied to ADNI without a matched-domain validation, and ADNI CN and AD scans may differ by site, scanner model, field strength, or acquisition parameters. Table 1 does not report site or scanner distributions, and the group t-tests in 'Comparing BAGs across cohorts' are unadjusted for these variables. The semi-global bias-correction coefficients are estimated on ADNI CN only and remove an average CA-related bias; they cannot remove AD-vs-CN technical shifts. Please report site/field-strength distributions and adjust or stratify the group comparison by site, or provide a sensitivity analysis restricted to a single scanner/field strength. Without this, the parahippocampal/temporal AD–CN difference map may reflect cohort composition rather than disease-specific aging.
  3. [Methods: Semi-global bias correction; Results: Comparing BAGs across cohorts] The semi-global correction averages vertex-specific slopes and intercepts into a single pair of coefficients and then applies the same adjustment to every vertex. This removes only a global CA-related trend and leaves vertex-specific bias intact. Since the AD–CN difference map is the main result, the sensitivity of the map to this choice should be quantified. A fully local correction, or a comparison between corrections, would indicate whether the reported parahippocampal difference (2.72 y) is robust or partly a residual artifact of the bias-correction procedure.
  4. [Results: BAGs predict cognitive scores] The regions used for cognitive regressions were selected post hoc: the parahippocampal gyrus was chosen because it showed the largest LBAG difference, and the orbital lateral sulcus because it showed the smallest. The reported p-values are therefore conditional on a selection made from the same data, and no correction for this selection is described. The claim that parahippocampal LBAGs are 'particularly strong' predictors of AD-related cognitive impairment is overstated unless the analysis is explicitly framed as exploratory or a selection-robust procedure is used (e.g., holding out region selection or correcting for the number of candidate regions considered).
minor comments (6)
  1. [Abstract; Methods: Data] The supplied abstract reports N = 14,423 for the training sample, whereas Methods and Table 1 report N = 14,250. Please reconcile this numerical inconsistency.
  2. [Methods: Statistical significance testing] The text says 'independent two-tailed t-tests' were used to test whether regional LBAs 'deviated significantly from each subject's CA.' Since each subject contributes both a regional LBA and a CA, this should be a paired/one-sample test on regional BAGs, not an independent-samples test. Please clarify the actual procedure.
  3. [Methods: Integrated gradients] For the AD group, the baseline is 'a randomly selected batch of CN participants.' Randomness in the baseline could affect saliency maps; please state whether the results are stable across baseline draws, or fix the baseline and report it.
  4. [Methods: Medial wall removal and smoothing] The smoothing procedure averages each node's LBA with neighbors up to two steps away and repeats this four times. Because the paper emphasizes high spatial resolution, it would be useful to state the effective spatial scale after smoothing or to provide a sensitivity analysis with different smoothing parameters.
  5. [Discussion: Technical novelty] The claim that 'no framework has been established to estimate LBA using cortical morphology' is too strong given the existence of voxel-based LBA models [26,27] and surface-based GBA models [35]. Consider softening to 'no LBA framework using cortical surface morphometry has been established' or similar.
  6. [Results: Feature contributions] The text says 'feature ablation highlights curvature and GWR as preferentially sensitive to AD pathology' in the abstract, but the Results only present integrated gradients. If ablation was actually performed, it should be described in Methods and reported; otherwise, the term 'ablation' should be removed.

Circularity Check

1 steps flagged · score 2.0 of 10

No load-bearing circularity; central GNN training/test chain is self-contained. One by-construction normalization (CN GBAG mean = 0) is reported as a result, and the post hoc region selection for cognitive regressions is a statistical concern, not a definitional circularity.

  1. self definitional [Methods – Semi-global bias correction; Results – Comparing BAGs across cohorts]
    "For each vertex, we regressed LBAGsv on CAs across subjects, obtaining a slope mv and intercept bv... We then averaged mv and bv across all v to obtain the semi-global slope mµ and intercept bµ... The corrected LBA per vertex and subject LBA′vs is thus defined as: LBA′vs = LBAvs − (mµCAs + bµ). After bias correction, GBAGs exhibited a mean of 0.00 y for the CNs."

    The semi-global correction coefficients are estimated by least-squares regression of LBAG on CA within the same ADNI CN cohort whose GBAG mean is then reported. Since the averaged slope and intercept are exactly the OLS coefficients for the vertex-averaged LBAG regressed on CA, subtracting them leaves a zero mean residual by construction; the 'mean = 0.00 y' is a mathematical identity, not an empirical result. This is a minor reporting tautology: it does not force the AD-CN spatial differences, which are computed after applying the same CN-derived correction to AD, so it is not load-bearing.

full rationale

The central derivation is not circular. The GNN is trained to predict vertex-level chronological age from cortical morphometry in CN adults and is evaluated on held-out ADNI scans; no fitted parameter is relabeled as a prediction. The semi-global bias correction is a standard post-hoc normalization, and its effect on the CN cohort (mean GBAG = 0) is by construction, but the paper's substantive findings - CN aging concentrated in prefrontal/parietal association cortex, AD-CN differences in parahippocampal and temporal regions, and cognitive associations - are not entailed by that normalization. The selection of the parahippocampal gyrus for cognitive regression because it had the largest LBAG difference is a post-hoc selection issue, not a definitional circularity, since the cognitive scores are external to the model. The paper's self-citations (e.g., Irimia et al. 2015, Amgalan et al. 2022) support background concepts and explainability methodology, but none is load-bearing in the way a uniqueness theorem or ansatz citation would be. The abstract's claim of lower MAE than the state of the art is internally inconsistent with the Discussion's statement that the model 'slightly (< 0.7 y) underperforms more recent variants,' but that is a correctness/consistency concern, not circularity. Overall, only a minor by-construction reporting of the CN zero mean justifies a small non-zero score.

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

No new physical or biological entities are postulated. LBA/LBAG are derived constructs, not new entities. The ledger instead captures the main fitted hyperparameters and domain assumptions on which the central claim rests.

free parameters (4)
  • Hidden feature sizes F1, F2 = F1=8, F2=16
    Chosen to yield lowest validation losses (Methods: Model structure).
  • Optimizer hyperparameters = learning rate=0.01, batch size=128, epochs=50
    Selected empirically (Methods: Model structure).
  • Smoothing parameters = 2-hop neighbors, 4 averaging iterations
    Chosen to remove visually apparent model artifacts (Methods: Medial wall removal and smoothing).
  • Semi-global bias-correction coefficients = m_mu and b_mu (not reported numerically)
    Per-vertex slopes/intercepts fit on ADNI CN and averaged; applied to AD (Methods: Semi-global bias correction).
assumptions (5)
  • domain assumption FreeSurfer-derived morphometric features (CT, SA, curvature, GWR, sulcal depth) represent biologically meaningful local aging signals.
    The entire input representation rests on this; no neuropathological or longitudinal ground truth validates it (Methods: Preprocessing).
  • domain assumption Chronological age is a valid supervised target for LBA, and residual LBAG after linear bias correction reflects biological age gap.
    Supervised training uses CA; residual interpretation is used throughout Results; no independent biological-age ground truth is available.
  • ad hoc to paper A model trained on CN adults from UKBB/NACC/IXI generalizes to ADNI AD without correcting for dataset/scanner/FreeSurfer-version shift.
    ADNI AD is a different cohort; disease effect is computed after applying CN-derived bias-correction coefficients (Methods: Semi-global bias correction).
  • ad hoc to paper Repeated 2-hop averaging (4 iterations) and atlas-based pooling preserve biological variation rather than erase it.
    Smoothing parameters were chosen to remove visually apparent artifacts (Methods: Medial wall removal and smoothing).
  • domain assumption Integrated gradients provide valid explanations of feature contributions.
    Authors acknowledge IGs share shortcomings of saliency methods (Limitations), yet feature-attribution conclusions are drawn from IGs.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Graph Neural Network Reveals the Cortical Morphology of Local Brain Aging in Normal Cognition and Alzheimer's Disease." pith.science (2026). https://pith.science/paper/P4H7Q5K5

@misc{pith2026260110912,
  author       = {Pith},
  title        = {Pith review of: Graph Neural Network Reveals the Cortical Morphology of Local Brain Aging in Normal Cognition and Alzheimer's Disease},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P4H7Q5K5}},
  note         = {Machine review of arXiv:2601.10912}
}
abstract

Estimating brain age (BA) from T1-weighted magnetic resonance images (MRIs) provides a powerful framework for quantifying anatomical brain aging. Whereas global BA (GBA) summarizes overall brain health, local BA (LBA) provides cortically specific patterns of aging at the subject level. Although previous studies have examined anatomical contributors to GBA, to our knowledge, no framework has been established to estimate LBA using cortical morphology. To address this gap, we introduce a graph neural network (GNN) that uses morphometric features$\unicode{x2013}$cortical thickness, surface area, curvature, gray/white matter intensity ratio (GWR), sulcal depth$\unicode{x2013}$to estimate LBA across the cortical surface at high spatial resolution (mean inter-vertex distance = 1.37 mm). Trained on cortical surface meshes extracted from the MRIs of cognitively normal (CN) adults (N = 14,423), our model achieves lower mean absolute error (MAE) than the existing state-of-the-art while identifying more biologically plausible patterns of aging in Alzheimer's disease (AD) on the ADNI dataset. Association cortices emerge as primary sites of morphometric aging in CNs, whereas mild cognitive impairment is characterized by widespread aging that is pronounced in the parahippocampal gyrus. AD subjects demonstrate significant aging across the entire cortex, particularly within medial temporal regions and associated cortical networks. Feature ablation highlights curvature and GWR as preferentially sensitive to AD pathology. Regional LBA gaps are significantly associated with neuropsychological measures of AD-related cognitive impairment, linking cortical aging patterns to clinical outcomes. These results demonstrate that GNN-based modeling of cortical morphometry enables biologically interpretable mapping of local brain aging with greater interpretability than prior work.

Figures

Figures reproduced from arXiv: 2601.10912 by the authors.

Figure 1
Figure 1. Graph U-Net architecture for cortical surface modeling. Blue rectangles denote feature maps defined on the cortical surface mesh at different atlas resolutions (e.g., V4, V5, V6). Joined blue and white rectangles represent skip connections, where features from earlier blocks are concatenated with those from later blocks. Each rectangle is annotated with the number of features F and vertices V, indicating mesh resolu… view at source ↗
Figure 2
Figure 2. Bias-corrected LBAGs for ADNI CNs and ADNI ADs – ADNI CNs (A) Bias-corrected LBAGs for CN subjects. (B) Vertex-wise differences between AD and CN LBAGs (AD minus CN). Feature contributions. Our saliency analysis using IGs revealed the local importance of each cortical feature in CNs ( [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. IGs for ADNI CN subjects. Group-level saliency maps showing the contribution of each cortical feature to model predictions: (A) SA, (B) CT, (C) GWR, (D) curvature, (E) sulcal depth. Saliency units retain relative significance across features [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Regression results between BAG and cognitive scores. [PITH_FULL_IMAGE:figures/full_fig_p016_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

132 extracted references · 8 linked inside Pith

  1. [27]

    A multitask deep learning model for voxel -level brain age estimation

    Gianchandani, N., et al. A multitask deep learning model for voxel -level brain age estimation . in International Workshop on Machine Learning in Medical Imaging. 2023. Springer

    2023

  2. [1]

    Transl Psychiatry, 2021

    Wu, B.S., et al., Cortical structure and the risk for Alzheimer's disease: a bidirectional Mendelian randomization study. Transl Psychiatry, 2021. 11(1): p. 476

    2021

  3. [2]

    Neuroimage Clin, 2023

    Bachmann, T., et al., Longitudinal changes in surface based brain morphometry measures in amnestic mild cognitive impairment and Alzheimer's Disease. Neuroimage Clin, 2023. 38 : p. 103371

    2023

  4. [3]

    Brain Res Bull, 2025

    Chen, L., et al., The impact of Alzheimer's disease on cortical complexity and its underlying biological mechanisms. Brain Res Bull, 2025. 225: p. 111320

    2025

  5. [4]

    Voelkle, and N

    Yuan, P., M.C. Voelkle, and N. Raz, Fluid intelligence and gross structural properties of the cerebral cortex in middle-aged and older adults: A multi-occasion longitudinal study. Neuroimage,

  6. [5]

    Machine vision and applications, 2013

    Dai, D., et al., Accurate prediction of AD patients using cortical thickness networks. Machine vision and applications, 2013. 24(7): p. 1445-1457

    2013

  7. [6]

    Andersson, and A

    Nyberg, L., M. Andersson, and A. Lundquist, Longitudinal change -change associations of cognition with cortical thickness and surface area. Aging Brain, 2023. 3: p. 100070

    2023

  8. [7]

    Cabral, and R.J

    Bauer, C.M., H.J. Cabral, and R.J. Killiany, Multimodal discrimination between normal aging, mild cognitive impairment and Alzheimer’s disease and prediction of cognitive decline. Diagnostics,

Show all 132 references
  1. [8]

    Neuroimage, 2018

    Shen, X., et al., Variation in longitudinal trajectories of cortical sulci in normal elderly. Neuroimage, 2018. 166: p. 1-9

    2018

  2. [9]

    J Alzheimers Dis, 2024

    Sighinolfi, G., et al., Sulcal Morphometry Predicts Mild Cognitive Impairment Conversion to Alzheimer's Disease. J Alzheimers Dis, 2024. 99(1): p. 177-190

    2024

  3. [10]

    Front Aging Neurosci, 2018

    Uribe, C., et al., Gray/White Matter Contrast in Parkinson's Disease. Front Aging Neurosci, 2018. 10: p. 89

    2018

  4. [11]

    Brain Imaging Behav, 2015

    Jefferson, A.L., et al., Gray & white matter tissue contrast differentiates Mild Cognitive Impairment converters from non-converters. Brain Imaging Behav, 2015. 9(2): p. 141-8

    2015

  5. [12]

    Front Aging Neurosci, 2021

    Lin, H.Y., et al., Differential Patterns of Gyral and Sulcal Morphological Changes During Normal Aging Process. Front Aging Neurosci, 2021. 13: p. 625931

    2021

  6. [13]

    Medicina (Kaunas), 2025

    Apse, R.R., et al., Morphometric Measurement of Mean Cortical Curvature: Analysis of Alterations in Cognitive Impairment. Medicina (Kaunas), 2025. 61(3)

    2025

  7. [14]

    J Neuropathol Exp Neurol, 2011

    Braak, H., et al., Stages of the pathologic process in Alzheimer disease: age categories from 1 to 100 years. J Neuropathol Exp Neurol, 2011. 70(11): p. 960-9

    2011

  8. [15]

    Brain Imaging Behav, 2015

    Irimia, A., et al., Statistical estimation of physiological brain age as a descriptor of senescence rate during adulthood. Brain Imaging Behav, 2015. 9(4): p. 678-89

    2015

  9. [16]

    Geroscience, 2022

    Amgalan, A., et al., Brain age estimation reveals older adults' accelerated senescence after traumatic brain injury. Geroscience, 2022. 44(5): p. 2509-2525

    2022

  10. [17]

    Neuroimage, 2010

    Franke, K., et al., Estimating the age of healthy subjects from T1-weighted MRI scans using kernel methods: exploring the influence of various parameters. Neuroimage, 2010. 50(3): p. 883-92

    2010

  11. [18]

    Molecular psychiatry, 2021

    Elliott, M.L., et al., Brain-age in midlife is associated with accelerated biological aging and cognitive decline in a longitudinal birth cohort. Molecular psychiatry, 2021. 26(8): p. 3829-3838

    2021

  12. [19]

    Proc Natl Acad Sci U S A, 2023

    Yin, C., et al., Anatomically interpretable deep learning of brain age captures domain- specific cognitive impairment. Proc Natl Acad Sci U S A, 2023. 120(2): p. e2214634120

    2023

  13. [20]

    Hum Brain Mapp,

    Kalc, P., et al., BrainAGE: Revisited and reframed machine learning workflow. Hum Brain Mapp,

  14. [21]

    Bioengineering (Basel), 2024

    Lim, H., et al., Brain Age Prediction Using Multi -Hop Graph Attention Combined with Convolutional Neural Network. Bioengineering (Basel), 2024. 11(3)

    2024

  15. [22]

    Ophthalmology, 2019

    Sayres, R., et al., Using a deep learning algorithm and integrated gradients explanation to assist grading for diabetic retinopathy. Ophthalmology, 2019. 126(4): p. 552-564

    2019

  16. [23]

    Radiology: Artificial Intelligence, 2021

    Arun, N., et al., Assessing the trustworthiness of saliency maps for localizing abnormalities in medical imaging. Radiology: Artificial Intelligence, 2021. 3(6): p. e200267

    2021

  17. [24]

    Radiology: Artificial Intelligence, 2023

    Zhang, J., et al., Revisiting the trustworthiness of saliency methods in radiology AI. Radiology: Artificial Intelligence, 2023. 6(1): p. e220221

    2023

  18. [25]

    Journal of Imaging Informatics in Medicine, 2024

    Venkatesh, K., et al., Gradient-Based Saliency Maps Are Not Trustworthy Visual Explanations of Automated AI Musculoskeletal Diagnoses. Journal of Imaging Informatics in Medicine, 2024. 37(5): p. 2490-2499

    2024

  19. [26]

    Front Aging Neurosci, 2021

    Popescu, S.G., et al., Local Brain-Age: A U-Net Model. Front Aging Neurosci, 2021. 13: p. 761954

    2021

  20. [28]

    Insights Imaging, 2024

    Gaeta, M., et al., T1 relaxation: Chemo-physical fundamentals of magnetic resonance imaging and clinical applications. Insights Imaging, 2024. 15(1): p. 200

    2024

  21. [29]

    IEEE Trans Med Imaging, 1999

    Van Leemput, K., et al., Automated model-based tissue classification of MR images of the brain. IEEE Trans Med Imaging, 1999. 18(10): p. 897-908

    1999

  22. [30]

    IEEE Transactions on Artificial Intelligence, 2021

    Rawal, A., et al., Recent advances in trustworthy explainable artificial intelligence: Status, challenges, and perspectives. IEEE Transactions on Artificial Intelligence, 2021. 3(6): p. 852-866

    2021

  23. [31]

    Geodesic convolutional neural networks on riemannian manifolds

    Masci, J., et al. Geodesic convolutional neural networks on riemannian manifolds. in Proceedings of the IEEE international conference on computer vision workshops. 2015

    2015

  24. [32]

    Spherical U -Net on cortical surfaces: methods and applications

    Zhao, F., et al. Spherical U -Net on cortical surfaces: methods and applications . in Information Processing in Medical Imaging: 26th International Conference, IPMI 2019, Hong Kong, China, June 2–7, 2019, Proceedings 26. 2019. Springer

    2019

  25. [33]

    Comparison between CNN and GNN pipelines for analysing the brain in development

    Bourlier, A., et al. Comparison between CNN and GNN pipelines for analysing the brain in development. in 20th International Conference on Computer Vision Theory and Applications. 2025. SCITEPRESS-Science and Technology Publications; SCITEPRESS

    2025

  26. [34]

    Sci Rep, 2025

    Tan, J., et al., Application of improved graph convolutional network for cortical surface parcellation. Sci Rep, 2025. 15(1): p. 16409

    2025

  27. [35]

    arXiv preprint arXiv:2411.05825, 2024

    Li, Z., et al., SurfGNN: A robust surface -based prediction model with interpretability for coactivation maps of spatial and cortical features. arXiv preprint arXiv:2411.05825, 2024

    2024

  28. [36]

    Medical image analysis, 2019

    Shin, S.Y., et al., Deep vessel segmentation by learning graphical connectivity. Medical image analysis, 2019. 58: p. 101556

    2019

  29. [37]

    arXiv preprint arXiv:2501.03765, 2025

    Singh, A., et al., Image Segmentation: Inducing graph- based learning. arXiv preprint arXiv:2501.03765, 2025

    2025 arXiv

  30. [38]

    Scientific Reports, 2025

    Xiao, H., et al., GNNs surpass transformers in tumor medical image segmentation. Scientific Reports, 2025. 15(1): p. 19842

    2025

  31. [39]

    Fischer, and T

    Ronneberger, O., P. Fischer, and T. Brox. U -net: Convolutional networks for biomedical image segmentation. in International Conference on Medical image computing and computer -assisted intervention. 2015. Springer

    2015

  32. [40]

    Alzheimer Dis Assoc Disord, 2007

    Beekly, D.L., et al., The National Alzheimer's Coordinating Center (NACC) database: the Uniform Data Set. Alzheimer Dis Assoc Disord, 2007. 21(3): p. 249-58

    2007

  33. [41]

    Neuroimage, 2016

    Kennedy, D.N., et al., The NITRC image repository. Neuroimage, 2016. 124(Pt B): p. 1069-1073

    2016

  34. [42]

    Nat Commun, 2020

    Littlejohns, T.J., et al., The UK Biobank imaging enhancement of 100,000 participants: rationale, data collection, management and future directions. Nat Commun, 2020. 11(1): p. 2624

    2020

  35. [43]

    Alzheimers Dement, 2013

    Weiner, M.W., et al., The Alzheimer's Disease Neuroimaging Initiative: a review of papers published since its inception. Alzheimers Dement, 2013. 9(5): p. e111-94

    2013

  36. [44]

    Fischl, and M.I

    Dale, A.M., B. Fischl, and M.I. Sereno, Cortical surface -based analysis. I. Segmentation and surface reconstruction. Neuroimage, 1999. 9(2): p. 179-94

    1999

  37. [45]

    Hum Brain Mapp, 1999

    Fischl, B., et al., High-resolution intersubject averaging and a coordinate system for the cortical surface. Hum Brain Mapp, 1999. 8(4): p. 272-84

    1999

  38. [46]

    arXiv preprint arXiv:1609.02907, 2016

    Kipf, T., Semi -supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907, 2016

    2016 arXiv

  39. [47]

    Ioffe, S. and C. Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. in International conference on machine learning. 2015. pmlr

    2015

  40. [48]

    arXiv preprint arXiv:1412.6980, 2014

    Kingma, D.P., Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014

    2014 arXiv

  41. [49]

    Neuroinformatics, 2024

    Guo, K.H., et al., Anatomic Interpretability in Neuroimage Deep Learning: Saliency Approaches for Typical Aging and Traumatic Brain Injury. Neuroinformatics, 2024. 22(4): p. 591-606

    2024

  42. [50]

    ArXiv, 2025

    Siegel, N.T., et al., Explainable AI Methods for Neuroimaging: Systematic Failures of Common Tools, the Need for Domain-Specific Validation, and a Proposal for Safe Application. ArXiv, 2025

    2025

  43. [51]

    Polsterl, and C

    Sarasua, I., S. Polsterl, and C. Wachinger, Hippocampal representations for deep learning on Alzheimer's disease. Sci Rep, 2022. 12(1): p. 8619

    2022

  44. [52]

    de Lange, A.G. and J.H. Cole, Commentary: Correction procedures in brain -age prediction. Neuroimage Clin, 2020. 26: p. 102229

    2020

  45. [53]

    Neuroimage Clin, 2019

    Beheshti, I., et al., Bias -adjustment in neuroimaging- based brain age frameworks: A robust scheme. Neuroimage Clin, 2019. 24: p. 102063

    2019

  46. [54]

    Neuroimage Clin, 2023

    Zhang, B., et al., Age-level bias correction in brain age prediction. Neuroimage Clin, 2023. 37: p. 103319

    2023

  47. [55]

    Cerebral Cortex, 2017

    Pehrs, C., et al., The temporal pole top- down modulates the ventral visual stream during social cognition. Cerebral Cortex, 2017. 27(1): p. 777-792

    2017

  48. [56]

    Navarro, and S

    Herlin, B., V. Navarro, and S. Dupont, The temporal pole: From anatomy to function-A literature appraisal. J Chem Neuroanat, 2021. 113: p. 101925

    2021

  49. [57]

    Front Psychiatry, 2022

    Guo, P., et al., Associations of Neurocognition and Social Cognition With Brain Structure and Function in Early-Onset Schizophrenia. Front Psychiatry, 2022. 13: p. 798105

    2022

  50. [58]

    Irimia, and I

    Rostowsky, K.A., A. Irimia, and I. Alzheimer's Disease Neuroimaging, Acute cognitive impairment after traumatic brain injury predicts the occurrence of brain atrophy patterns similar to those observed in Alzheimer's disease. Geroscience, 2021. 43(4): p. 2015-2039

    2021

  51. [59]

    Brain Topogr, 2011

    McGinnis, S.M., et al., Age -related changes in the thickness of cortical zones in humans. Brain Topogr, 2011. 24(3-4): p. 279-91

    2011

  52. [60]

    Cereb Cortex, 2013

    Hogstrom, L.J., et al., The structure of the cerebral cortex across adult life: age -related patterns of surface area, thickness, and gyrification. Cereb Cortex, 2013. 23(11): p. 2521-30

    2013

  53. [61]

    Eur J Neurosci, 2021

    Madan, C.R., Age -related decrements in cortical gyrification: Evidence from an accelerated longitudinal dataset. Eur J Neurosci, 2021. 53(5): p. 1661-1671

    2021

  54. [62]

    Proc Natl Acad Sci U S A, 2014

    Douaud, G., et al., A common brain network links development, aging, and vulnerability to disease. Proc Natl Acad Sci U S A, 2014. 111(49): p. 17648-53

    2014

  55. [63]

    Wandell, and A.A

    Yeatman, J.D., B.A. Wandell, and A.A. Mezer, Lifespan maturation and degeneration of human brain white matter. Nat Commun, 2014. 5: p. 4932

    2014

  56. [64]

    Proc Natl Acad Sci U S A, 2010

    Hill, J., et al., Similar patterns of cortical expansion during human development and evolution. Proc Natl Acad Sci U S A, 2010. 107(29): p. 13135-40

    2010

  57. [65]

    Neuroimage, 2019

    Dubois, J., et al., The dynamics of cortical folding waves and prematurity- related deviations revealed by spatial and spectral analysis of gyrification. Neuroimage, 2019. 185: p. 934-946

    2019

  58. [66]

    Nature Communications, 2021

    Voorhies, W.I., et al., Cognitive insights from tertiary sulci in prefrontal cortex. Nature Communications, 2021. 12(1): p. 5122

    2021

  59. [67]

    Ageing Res Rev, 2016

    Pini, L., et al., Brain atrophy in Alzheimer's Disease and aging. Ageing Res Rev, 2016. 30: p. 25- 48

    2016

  60. [68]

    J Neurosci, 2022

    de Flores, R., et al., Medial Temporal Lobe Networks in Alzheimer's Disease: Structural and Molecular Vulnerabilities. J Neurosci, 2022. 42(10): p. 2131-2141

    2022

  61. [69]

    Migliaccio, R. and F. Cacciamani, The temporal lobe in typical and atypical Alzheimer disease. Handb Clin Neurol, 2022. 187: p. 449-466

    2022

  62. [70]

    Morris, and B.C

    Bakkour, A., J.C. Morris, and B.C. Dickerson, The cortical signature of prodromal AD: regional thinning predicts mild AD dementia. Neurology, 2009. 72(12): p. 1048-55

    2009

  63. [71]

    Cereb Cortex, 2009

    Dickerson, B.C., et al., The cortical signature of Alzheimer's disease: regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals. Cereb Cortex, 2009. 19(3): p. 497-510

    2009

  64. [72]

    Alzheimers Dement (Amst),

    Verfaillie, S.C., et al., Thinner temporal and parietal cortex is related to incident clinical progression to dementia in patients with subjective cognitive decline. Alzheimers Dement (Amst),

  65. [73]

    Klobusiakova, and I

    Krajcovicova, L., P. Klobusiakova, and I. Rektorova, Gray Matter Changes in Parkinson's and Alzheimer's Disease and Relation to Cognition. Curr Neurol Neurosci Rep, 2019. 19(11): p. 85

    2019

  66. [74]

    bioRxiv, 2025: p

    Dzianok, P., et al., Cortical thinning in temporal pole, a core region in Alzheimer’s disease, in non- demented, middle -aged APOE -ε4 and PICALM -AA/AG carriers. bioRxiv, 2025: p. 2025.02. 21.639542

    2025

  67. [75]

    Journal of Neuroscience, 2012

    Libby, L.A., et al., Differential connectivity of perirhinal and parahippocampal cortices within human hippocampal subregions revealed by high- resolution functional imaging. Journal of Neuroscience, 2012. 32(19): p. 6550-6560

    2012

  68. [76]

    Sci Rep, 2019

    Maller, J.J., et al., Revealing the Hippocampal Connectome through Super -Resolution 1150- Direction Diffusion MRI. Sci Rep, 2019. 9(1): p. 2418

    2019

  69. [77]

    Kveraga, and M

    Aminoff, E.M., K. Kveraga, and M. Bar, The role of the parahippocampal cortex in cognition. Trends Cogn Sci, 2013. 17(8): p. 379-90

    2013

  70. [78]

    Association., A.P., Diagnostic and statistical manual of mental disorders (5th ed.) . 5 ed. 2013: American Psychiatric Association Publishing

    2013

  71. [79]

    J Alzheimers Dis, 2011

    Scheff, S.W., et al., Synaptic loss in the inferior temporal gyrus in mild cognitive impairment and Alzheimer's disease. J Alzheimers Dis, 2011. 24(3): p. 547-57

    2011

  72. [80]

    Basic Clin Neurosci, 2022

    Sattari, N., et al., Assessing the Changes of Cortical Thickness in Alzheimer Disease With MRI Using Freesurfer Software. Basic Clin Neurosci, 2022. 13(2): p. 185-192

    2022

  73. [81]

    Alzheimers Dement, 2025

    Wuestefeld, A., et al., Tau, atrophy, and domain -specific cognitive impairment in typical Alzheimer's disease. Alzheimers Dement, 2025. 21(7): p. e70511

    2025

  74. [82]

    Neurobiology of aging, 2000

    Convit, A., et al., Atrophy of the medial occipitotemporal, inferior, and middle temporal gyri in non-demented elderly predict decline to Alzheimer’s disease ☆. Neurobiology of aging, 2000. 21(1): p. 19-26

    2000

  75. [83]

    J Integr Neurosci,

    Lu, F., et al., The Correlations between Volume Loss of Temporal and Subcortical Functional Subregions and Cognitive Impairment at Various Stages of Cognitive Decline. J Integr Neurosci,

  76. [84]

    Cereb Cortex,

    Arnold, S.E., et al., The topographical and neuroanatomical distribution of neurofibrillary tangles and neuritic plaques in the cerebral cortex of patients with Alzheimer's disease. Cereb Cortex,

  77. [85]

    Hyman, and G.W

    Arnold, S.E., B.T. Hyman, and G.W. Van Hoesen, Neuropathologic changes of the temporal pole in Alzheimer's disease and Pick's disease. Arch Neurol, 1994. 51(2): p. 145-50

    1994

  78. [86]

    Neuroimage, 2017

    LaPoint, M.R., et al., The association between tau PET and retrospective cortical thinning in clinically normal elderly. Neuroimage, 2017. 157: p. 612-622

    2017

  79. [87]

    PLoS One, 2017

    Rasero, J., et al., Multivariate regression analysis of structural MRI connectivity matrices in Alzheimer’s disease. PLoS One, 2017. 12(11): p. e0187281

    2017

  80. [88]

    Brain Sciences, 2021

    Lenhart, L., et al., Anatomically Standardized Detection of MRI Atrophy Patterns in Early -Stage Alzheimer’s Disease. Brain Sciences, 2021. 11(11): p. 1491

    2021

  81. [89]

    Imaging Neuroscience, 2025

    Stocks, J., et al., Spatial and Temporal Progression of Neurodegeneration in Confirmed and Suspected TDP-43 Type C Pathology. Imaging Neuroscience, 2025

    2025

  82. [90]

    General psychiatry, 2019

    Yang, H., et al., Study of brain morphology change in Alzheimer’s disease and amnestic mild cognitive impairment compared with normal controls. General psychiatry, 2019. 32(2): p. e100005

    2019

  83. [91]

    Neurobiology of aging, 2003

    Halliday, G., et al., Identifying severely atrophic cortical subregions in Alzheimer’s disease. Neurobiology of aging, 2003. 24(6): p. 797-806

    2003

  84. [92]

    Neurobiology of aging, 1997

    Convit, A., et al., Specific hippocampal volume reductions in individuals at risk for Alzheimer’s disease. Neurobiology of aging, 1997. 18(2): p. 131-138

    1997

  85. [93]

    Clinical epigenetics, 2020

    Ma, D., et al., The fusiform gyrus exhibits an epigenetic signature for Alzheimer’s disease. Clinical epigenetics, 2020. 12(1): p. 129

    2020

  86. [94]

    Predicting progression from mild cognitive impairment to Alzheimer’s disease using MRI-based cortical features and a two-state Markov Model

    Ficiarà, E., et al. Predicting progression from mild cognitive impairment to Alzheimer’s disease using MRI-based cortical features and a two-state Markov Model. in 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI). 2021. IEEE

    2021

  87. [95]

    Neurobiology of Aging, 2022

    Mortamais, M., et al., Sulcal morphology as cognitive decline predictor in older adults with memory complaints. Neurobiology of Aging, 2022. 113: p. 84-94

    2022

  88. [96]

    Geroscience, 2024

    Yu, J., Age-related decline in thickness and surface area in the cortical surface and hippocampus: lifespan trajectories and decade-by-decade analyses. Geroscience, 2024. 46(6): p. 6213-6227

    2024

  89. [97]

    J Neurosci, 2008

    Bassett, D.S., et al., Hierarchical organization of human cortical networks in health and schizophrenia. J Neurosci, 2008. 28(37): p. 9239-48

    2008

  90. [98]

    He, Y. and A. Evans, Graph theoretical modeling of brain connectivity. Curr Opin Neurol, 2010. 23(4): p. 341-50

    2010

  91. [99]

    Neuroimage, 2010

    Liu, T., et al., The effects of age and sex on cortical sulci in the elderly. Neuroimage, 2010. 51(1): p. 19-27

    2010

  92. [100]

    Ronan, L. and P.C. Fletcher, From genes to folds: a review of cortical gyrification theory. Brain Struct Funct, 2015. 220(5): p. 2475-83

    2015

  93. [101]

    Proc Natl Acad Sci U S A, 2020

    Alexander-Bloch, A.F., et al., Imaging local genetic influences on cortical folding. Proc Natl Acad Sci U S A, 2020. 117(13): p. 7430-7436

    2020

  94. [102]

    Exposito-Alonso, and C.A

    Akula, S.K., D. Exposito-Alonso, and C.A. Walsh, Shaping the brain: The emergence of cortical structure and folding. Dev Cell, 2023. 58(24): p. 2836-2849

    2023

  95. [103]

    Neuroimage, 2023

    Huang, Y., et al., Genetic Influence on Gyral Peaks. Neuroimage, 2023. 280: p. 120344

    2023

  96. [104]

    Brain Struct Funct, 2025

    Cao, G., et al., Gyral peak variations between HCP and CHCP: functional and structural implications. Brain Struct Funct, 2025. 230(2): p. 37

    2025

  97. [105]

    Hum Brain Mapp, 2022

    Zhang, S., et al., Gyral peaks: Novel gyral landmarks in developing macaque brains. Hum Brain Mapp, 2022. 43(15): p. 4540-4555

    2022

  98. [106]

    Cereb Cortex, 2023

    Zhang, S., et al., Gyral peaks and patterns in human brains. Cereb Cortex, 2023. 33(11): p. 6708- 6722

    2023

  99. [107]

    Cereb Cortex, 2009

    Panizzon, M.S., et al., Distinct genetic influences on cortical surface area and cortical thickness. Cereb Cortex, 2009. 19(11): p. 2728-35

    2009

  100. [108]

    Science, 2020

    Grasby, K.L., et al., The genetic architecture of the human cerebral cortex. Science, 2020. 367(6484)

    2020

  101. [109]

    Quantitative Imaging in Medicine and Surgery, 2022

    Cui, D., et al., Age-and sex-related differences in cortical morphology and their relationships with cognitive performance in healthy middle-aged and older adults. Quantitative Imaging in Medicine and Surgery, 2022. 13(2): p. 1083

    2022

  102. [110]

    Neuroinformatics, 2021

    Irimia, A., Cross- Sectional Volumes and Trajectories of the Human Brain, Gray Matter, White Matter and Cerebrospinal Fluid in 9473 Typically Aging Adults. Neuroinformatics, 2021. 19(2): p. 347-366

    2021

  103. [111]

    Neurosci Lett, 2013

    van Velsen, E.F., et al., Brain cortical thickness in the general elderly population: the Rotterdam Scan Study. Neurosci Lett, 2013. 550: p. 189-94

    2013

  104. [112]

    J Neurosci,

    Storsve, A.B., et al., Differential longitudinal changes in cortical thickness, surface area and volume across the adult life span: regions of accelerating and decelerating change. J Neurosci,

  105. [113]

    Neuroimage, 2011

    Salat, D.H., et al., Hippocampal degeneration is associated with temporal and limbic gray matter/white matter tissue contrast in Alzheimer's disease. Neuroimage, 2011. 54(3): p. 1795-1802

    2011

  106. [114]

    NeuroImage: Clinical, 2023

    Putcha, D., et al., Gray to white matter signal ratio as a novel biomarker of neurodegeneration in Alzheimer’s disease. NeuroImage: Clinical, 2023. 37: p. 103303

    2023

  107. [115]

    Neuroimage, 2009

    Salat, D.H., et al., Age-associated alterations in cortical gray and white matter signal intensity and gray to white matter contrast. Neuroimage, 2009. 48(1): p. 21-8

    2009

  108. [116]

    PLoS One,

    Kong, L., et al., Reduced gray to white matter tissue intensity contrast in schizophrenia. PLoS One,

  109. [117]

    Proceedings of the National Academy of Sciences, 2020

    Van Essen, D.C., A 2020 view of tension -based cortical morphogenesis. Proceedings of the National Academy of Sciences, 2020. 117(52): p. 32868-32879

    2020

  110. [118]

    J Neurosci, 2024

    Maboudian, S.A., et al., Defining Overlooked Structures Reveals New Associations between the Cortex and Cognition in Aging and Alzheimer's Disease. J Neurosci, 2024. 44(16)

    2024

  111. [119]

    Neuroimage Clin, 2014

    Deppe, M., et al., Increased cortical curvature reflects white matter atrophy in individual patients with early multiple sclerosis. Neuroimage Clin, 2014. 6: p. 475-87

    2014

  112. [120]

    Demirci, N. and M.A. Holland, Cortical thickness systematically varies with curvature and depth in healthy human brains. Hum Brain Mapp, 2022. 43(6): p. 2064-2084

    2022

  113. [121]

    Alzheimer Disease & Associated Disorders, 2010

    Teng, E., et al., Utility of the functional activities questionnaire for distinguishing mild cognitive impairment from very mild Alzheimer disease. Alzheimer Disease & Associated Disorders, 2010. 24(4): p. 348-353

    2010

  114. [122]

    Alzheimer's & Dementia, 2011

    Coley, N., et al., Suitability of the Clinical Dementia Rating- Sum of Boxes as a single primary endpoint for Alzheimer’s disease trials. Alzheimer's & Dementia, 2011. 7(6): p. 602-610. e2

    2011

  115. [123]

    Speechley, and M

    Kueper, J.K., M. Speechley, and M. Montero-Odasso, The Alzheimer’s disease assessment scale– cognitive subscale (ADAS-Cog): modifications and responsiveness in pre-dementia populations. a narrative review. Journal of Alzheimer’s Disease, 2018. 63(2): p. 423-444

    2018

  116. [124]

    American Journal of Alzheimer's Disease & Other Dementias®, 2020

    Benoit, J.S., et al., Longitudinal sensitivity of Alzheimer’s disease severity staging. American Journal of Alzheimer's Disease & Other Dementias®, 2020. 35: p. 1533317520918719

    2020

  117. [125]

    Alon, U. and E. Yahav, On the bottleneck of graph neural networks and its practical implications. arXiv preprint arXiv:2006.05205, 2020

    2006 arXiv

  118. [126]

    Lackenby, and P

    Deac, A., M. Lackenby, and P. Veličković. Expander graph propagation. in Learning on Graphs Conference. 2022. PMLR

    2022

  119. [127]

    Bechler -Speicher, and P

    Wilson, J., M. Bechler -Speicher, and P. Veličković, Cayley graph propagation. arXiv preprint arXiv:2410.03424, 2024

    2024 arXiv

  120. [128]

    arXiv preprint arXiv:2310.01668, 2023

    Barbero, F., et al., Locality-aware graph-rewiring in gnns. arXiv preprint arXiv:2310.01668, 2023

    2023 arXiv

  121. [129]

    Buscaldi, and N

    Attali, H., D. Buscaldi, and N. Pernelle, Rewiring Techniques to Mitigate Oversquashing and Oversmoothing in GNNs: A Survey. arXiv preprint arXiv:2411.17429, 2024

    2024 arXiv

  122. [130]

    Neural message passing for quantum chemistry

    Gilmer, J., et al. Neural message passing for quantum chemistry . in International conference on machine learning. 2017. PMLR

    2017

  123. [131]

    arXiv preprint arXiv:2405.17311, 2024

    Qian, C., et al., Probabilistic graph rewiring via virtual nodes. arXiv preprint arXiv:2405.17311, 2024

    2024 arXiv

  124. [132]

    Scientific reports, 2019

    Ni, C.-C., et al., Community detection on networks with Ricci flow. Scientific reports, 2019. 9(1): p. 9984. Supplementary Figures Supplementary Figure 1. Distribution of GBAGs for the ADNI test sets. (A) Distribution of GBAGs for the ADNI CN and ADNI AD test sets. Dashed line...

    2019

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.