Pith. sign in

REVIEW 2 major objections 2 minor 47 references

A model pretrained once on labeling brain streamlines produces embeddings that parcellate tracts and predict age and sex across new datasets and methods without retraining.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

TractFM is a new tractogram foundation model pretrained on anatomical streamline parcellation that produces transferable streamline and subject-level representations across multiple dMRI datasets and tractography methods.

T0 review reviewed 2026-06-28 challenge →

load-bearing objection TractFM combines a local streamline encoder with a permutation-equivariant tractogram encoder and pretrains on anatomical parcellation to produce reusable embeddings for both local labeling and subject phenotype prediction. the 2 major comments →

arxiv 2606.09893 v1 pith:55GOECTR submitted 2026-06-03 eess.IV cs.AIcs.LG

Tractogram foundation model

Guikun Chen , Yuqian Chen , Yijie Li , Yogesh Rathi , Nikos Makris , Fan Zhang , Wenguan Wang , Lauren J. O'Donnell This is my paper
classification eess.IV cs.AIcs.LG
keywords tractographydiffusion MRIfoundation modelwhite matterstreamline classificationrepresentation learningbrain connectivitypermutation equivariance
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

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 presents TractFM as a foundation model that processes entire tractograms from diffusion MRI in one pass. It pretrains by assigning anatomical labels to streamlines while learning joint geometric context across all streamlines in a subject. This produces both per-streamline embeddings for tract mapping and compact subject descriptors for phenotype prediction. The frozen representations transfer to new tractography pipelines and independent datasets for accurate parcellation and for predicting age and sex. A reader would care if one set of learned whole-brain patterns can replace separate hand-crafted features or task-specific models for multiple uses in brain connectivity analysis.

Core claim

TractFM combines a local streamline encoder with a permutation-equivariant tractogram encoder. Pretraining on dense anatomical tract parcellation produces contextualized streamline-level embeddings for tract parcellation and compact subject-level descriptors for downstream prediction of subject phenotypes. These representations transfer across three tractography algorithms and five dMRI datasets while remaining frozen, showing that whole-brain geometric context learned once generalizes to streamline-level and subject-level tasks.

What carries the argument

The permutation-equivariant tractogram encoder that jointly contextualizes all streamlines from one subject during pretraining on anatomical label assignment.

Load-bearing premise

Pretraining solely on assigning anatomical labels to individual streamlines produces embeddings whose subject-level summaries capture phenotype-relevant variation without requiring task-specific fine-tuning or additional supervision.

What would settle it

An experiment on a sixth unseen dMRI dataset and fourth tractography method where the frozen TractFM embeddings yield tract parcellation accuracy below current separate classifiers or show no statistically significant correlation with age or sex.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Frozen streamline embeddings achieve accurate tract parcellation when applied to tractograms from different algorithms and scanners.
  • Subject-level descriptors derived from the same pretraining predict age and sex on independent datasets without further training.
  • The model unifies local streamline geometry with whole-brain anatomical organization in reusable representations.
  • Representations connect streamline classification directly to subject-level phenotype prediction in a single pretrained network.

Where Pith is reading between the lines

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

  • If the learned context proves stable, future studies could test whether the same descriptors correlate with clinical variables such as cognitive scores or disease status.
  • The approach might allow researchers to apply one set of embeddings to new prediction tasks instead of redesigning features for each study.
  • Extending pretraining to include more varied acquisition protocols could further test whether the geometric patterns remain useful outside the current five datasets.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 2 minor

Summary. The paper introduces TractFM, a tractogram foundation model that combines a local streamline encoder with a permutation-equivariant tractogram encoder. Pretrained on dense anatomical tract parcellation (assigning labels to individual streamlines), it produces contextualized streamline-level embeddings and compact subject-level descriptors. These are claimed to transfer without fine-tuning to accurate tract parcellation and to prediction of age and sex, generalizing across three tractography algorithms and five dMRI datasets.

Significance. If the central claims are supported by quantitative evidence, the work would be significant for representation learning in neuroimaging: it offers a single pretraining procedure that yields reusable representations bridging local streamline geometry and whole-brain inter-subject variation, moving beyond separate hand-crafted or task-specific approaches.

major comments (2)
  1. [Methods (pretraining objective)] Methods (pretraining objective description): The pretraining loss is defined solely on per-streamline anatomical label assignment. No auxiliary objective, contrastive term, or regularization is stated that would force the aggregated subject-level descriptor (output of the permutation-equivariant encoder) to retain geometric variation orthogonal to label statistics. This directly engages the load-bearing assumption that phenotype prediction reflects learned whole-brain context rather than recoverable label-count confounds.
  2. [Results] Results section: The abstract states that frozen representations achieve accurate tract parcellation and predict age and sex across three algorithms and five datasets, yet the manuscript supplies no numerical performance values, error bars, ablation tables, or statistical tests. Without these, the generalization claim cannot be evaluated and the transfer results remain unverified.
minor comments (2)
  1. [Abstract] Abstract: The single long paragraph mixes model architecture, pretraining task, and transfer claims; breaking it into two paragraphs would improve readability.
  2. [Notation / Model architecture] Notation: The manuscript would benefit from explicit symbols distinguishing the per-streamline embedding from the subject-level summary vector, together with a schematic of the two-stage encoder.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments, which help clarify the presentation of our work. We address each major comment below and outline the corresponding revisions.

read point-by-point responses

  1. Referee: Methods (pretraining objective description): The pretraining loss is defined solely on per-streamline anatomical label assignment. No auxiliary objective, contrastive term, or regularization is stated that would force the aggregated subject-level descriptor (output of the permutation-equivariant encoder) to retain geometric variation orthogonal to label statistics. This directly engages the load-bearing assumption that phenotype prediction reflects learned whole-brain context rather than recoverable label-count confounds.

    Authors: The pretraining objective is the per-streamline anatomical label prediction loss. The permutation-equivariant encoder is intended to enable joint contextualization across the entire tractogram, so that the resulting subject-level descriptor encodes inter-streamline geometric relationships in addition to label information. We agree that the manuscript does not describe an auxiliary loss or regularization term that explicitly enforces orthogonality to label-count statistics. In the revised manuscript we will add a targeted analysis (e.g., comparison against a label-histogram baseline and partial-correlation controls) to quantify the contribution of geometric context beyond label counts. revision: yes

  2. Referee: Results section: The abstract states that frozen representations achieve accurate tract parcellation and predict age and sex across three algorithms and five datasets, yet the manuscript supplies no numerical performance values, error bars, ablation tables, or statistical tests. Without these, the generalization claim cannot be evaluated and the transfer results remain unverified.

    Authors: The current manuscript version indeed omits explicit numerical tables, error bars, ablation results, and statistical tests from the main Results section. We will revise the manuscript to include these quantitative elements (performance metrics with standard deviations across folds, ablation tables, and appropriate statistical tests) directly in the main text so that the generalization claims can be evaluated. revision: yes

Circularity Check

0 steps flagged

No circularity: pretraining objective is distinct from transfer tasks with no equations reducing outputs to inputs

full rationale

The paper presents a standard pretrain-then-transfer setup: pretraining optimizes streamline anatomical labeling, after which frozen embeddings are evaluated on separate parcellation and phenotype prediction tasks across datasets. No equations, fitted parameters renamed as predictions, or self-citation chains are described that would make the reported generalization equivalent to the pretraining inputs by construction. The central claim rests on empirical transfer performance rather than any definitional reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review yields no explicit free parameters, axioms, or invented entities beyond the model itself; the central claim rests on the unstated assumption that the described encoder architecture and pretraining objective are sufficient to produce the claimed generalization.

reviewed 2026-06-28 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Tractogram foundation model." pith.science (2026). https://pith.science/paper/55GOECTR

@misc{pith2026260609893,
  author       = {Pith},
  title        = {Pith review of: Tractogram foundation model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/55GOECTR}},
  note         = {Machine review of arXiv:2606.09893}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Diffusion MRI (dMRI) tractography is the only noninvasive approach for mapping white-matter pathways in the living human brain. It represents each brain as a tractogram: a large, unordered set of three-dimensional streamlines that includes information about both local streamline geometry and whole-brain anatomical organization. This structure makes tractograms a natural but challenging target for representation learning. Existing methods treat streamline classification and subject-level prediction as separate problems: streamline classifiers focus on geometric patterns, whereas subject-level prediction often depends on hand-crafted features. As a result, current methods do not learn reusable representations that connect streamline anatomy with whole-brain inter-subject variation. Here we introduce TractFM, a tractogram foundation model that learns reusable representations directly from whole-brain streamline sets. TractFM combines a local streamline encoder with a permutation-equivariant tractogram encoder, allowing all streamlines from a subject to be contextualized jointly in a single forward pass. Pretraining on dense anatomical tract parcellation, i.e., assigning anatomical labels to individual streamlines, yields two complementary representations: contextualized streamline-level embeddings for tract parcellation and compact subject-level descriptors for downstream prediction of subject phenotypes. Across three tractography algorithms and five dMRI datasets, TractFM transfers to both streamline-level and subject-level tasks. Its frozen representations achieve accurate tract parcellation and predict age and sex across independent datasets. These results show that whole-brain geometric context, learned once, can generalize across tractography pipelines, datasets, and prediction tasks.

Figures

Figures reproduced from arXiv: 2606.09893 by Fan Zhang, Guikun Chen, Lauren J. O'Donnell, Nikos Makris, Wenguan Wang, Yijie Li, Yogesh Rathi, Yuqian Chen.

Figure 1
Figure 1. Figure 1: Overview of the TractFM whole-brain tractogram modeling framework. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: t-SNE visualization of frozen contextualized streamline embeddings learned by TractFM. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Selective attention profile of TractFM for a representative corpus callosum streamline. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

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

47 extracted references

  1. [1]

    Diffusion mri fiber tractography of the brain.NMR in Biomedicine, 32(4):e3785, 2019

    Ben Jeurissen, Maxime Descoteaux, Susumu Mori, and Alexander Leemans. Diffusion mri fiber tractography of the brain.NMR in Biomedicine, 32(4):e3785, 2019

    2019

  2. [2]

    Quantitative mapping of the brain’s structural connectivity using diffusion mri tractography: A review.Neuroimage, 249:118870, 2022

    Fan Zhang, Alessandro Daducci, Yong He, Simona Schiavi, Caio Seguin, Robert E Smith, Chun-Hung Yeh, Tengda Zhao, and Lauren J O’Donnell. Quantitative mapping of the brain’s structural connectivity using diffusion mri tractography: A review.Neuroimage, 249:118870, 2022

    2022

  3. [3]

    Mapping the structural core of human cerebral cortex.PLoS biology, 6(7):e159, 2008

    Patric Hagmann, Leila Cammoun, Xavier Gigandet, Reto Meuli, Christopher J Honey, Van J Wedeen, and Olaf Sporns. Mapping the structural core of human cerebral cortex.PLoS biology, 6(7):e159, 2008

    2008

  4. [4]

    Anatomically-constrained tractography: improved diffusion mri streamlines tractography through effective use of anatomical information.Neuroimage, 62(3):1924–1938, 2012

    Robert E Smith, Jacques-Donald Tournier, Fernando Calamante, and Alan Connelly. Anatomically-constrained tractography: improved diffusion mri streamlines tractography through effective use of anatomical information.Neuroimage, 62(3):1924–1938, 2012

    1924

  5. [5]

    Anatomical accuracy of brain connections derived from diffu- sion mri tractography is inherently limited.Proceedings of the National Academy of Sciences, 111(46):16574–16579, 2014

    Cibu Thomas, Frank Q Ye, M Okan Irfanoglu, Pooja Modi, Kadharbatcha S Saleem, David A Leopold, and Carlo Pierpaoli. Anatomical accuracy of brain connections derived from diffu- sion mri tractography is inherently limited.Proceedings of the National Academy of Sciences, 111(46):16574–16579, 2014

    2014

  6. [6]

    Potential and limitations of diffusion mri tractogra- phy for the study of language.Brain and language, 131:65–73, 2014

    Jennifer SW Campbell and G Bruce Pike. Potential and limitations of diffusion mri tractogra- phy for the study of language.Brain and language, 131:65–73, 2014. 16

    2014

  7. [7]

    A diffusion mri tractography connectome of the mouse brain and comparison with neuronal tracer data

    Evan Calabrese, Alexandra Badea, Gary Cofer, Yi Qi, and G Allan Johnson. A diffusion mri tractography connectome of the mouse brain and comparison with neuronal tracer data. Cerebral cortex, 25(11):4628–4637, 2015

    2015

  8. [8]

    Brain connectomics predict response to treatment in social anxiety disorder.Molecular psychiatry, 21(5):680–685, 2016

    S Whitfield-Gabrieli, SS Ghosh, A Nieto-Castanon, Z Saygin, O Doehrmann, XJ Chai, GO Reynolds, SG Hofmann, MH Pollack, and JDE Gabrieli. Brain connectomics predict response to treatment in social anxiety disorder.Molecular psychiatry, 21(5):680–685, 2016

    2016

  9. [9]

    The challenge of mapping the human connectome based on diffusion tractography.Nature communications, 8(1):1349, 2017

    Klaus H Maier-Hein, Peter F Neher, Jean-Christophe Houde, Marc-Alexandre Côté, Eleftherios Garyfallidis, Jidan Zhong, Maxime Chamberland, Fang-Cheng Yeh, Ying-Chia Lin, Qing Ji, et al. The challenge of mapping the human connectome based on diffusion tractography.Nature communications, 8(1):1349, 2017

    2017

  10. [10]

    Mapping connectomes with diffusion mri: deterministic or probabilistic tractography?Magnetic resonance in medicine, 81(2):1368–1384, 2019

    Tabinda Sarwar, Kotagiri Ramamohanarao, and Andrew Zalesky. Mapping connectomes with diffusion mri: deterministic or probabilistic tractography?Magnetic resonance in medicine, 81(2):1368–1384, 2019

    2019

  11. [11]

    Limits to anatomicalaccuracyofdiffusiontractographyusing modernapproaches.Neuroimage, 185:1–11, 2019

    Kurt G Schilling, Vishwesh Nath, Colin Hansen, Prasanna Parvathaneni, Justin Blaber, Yurui Gao, Peter Neher, Dogu Baran Aydogan, Yonggang Shi, Mario Ocampo-Pineda, et al. Limits to anatomicalaccuracyofdiffusiontractographyusing modernapproaches.Neuroimage, 185:1–11, 2019

    2019

  12. [12]

    Mapping structural connectivity using diffusion mri: Challenges and opportunities.Journal of Magnetic Resonance Imaging, 53(6):1666–1682, 2021

    Chun-Hung Yeh, Derek K Jones, Xiaoyun Liang, Maxime Descoteaux, and Alan Connelly. Mapping structural connectivity using diffusion mri: Challenges and opportunities.Journal of Magnetic Resonance Imaging, 53(6):1666–1682, 2021

    2021

  13. [13]

    Think deep in the tractography game: deep learning for tractography comput- ing and analysis.Brain Structure and Function, 230(6):100, 2025

    Fan Zhang, Antoine Théberge, Pierre-Marc Jodoin, Maxime Descoteaux, and Lauren J O’Donnell. Think deep in the tractography game: deep learning for tractography comput- ing and analysis.Brain Structure and Function, 230(6):100, 2025

    2025

  14. [14]

    Foundation models for generalist medical artificial intelligence.Nature, 616(7956):259–265, 2023

    Michael Moor, Oishi Banerjee, Zahra Shakeri Hossein Abad, Harlan M Krumholz, Jure Leskovec, Eric J Topol, and Pranav Rajpurkar. Foundation models for generalist medical artificial intelligence.Nature, 616(7956):259–265, 2023

    2023

  15. [15]

    Foundation models in radiology: what, how, why, and why not.Radiology, 314(2):e240597, 2025

    Magdalini Paschali, Zhihong Chen, Louis Blankemeier, Maya Varma, Alaa Youssef, Christian Bluethgen, Curtis Langlotz, Sergios Gatidis, and Akshay Chaudhari. Foundation models in radiology: what, how, why, and why not.Radiology, 314(2):e240597, 2025

    2025

  16. [16]

    Foun- dation model for cancer imaging biomarkers.Nature machine intelligence, 6(3):354–367, 2024

    Suraj Pai, Dennis Bontempi, Ibrahim Hadzic, Vasco Prudente, Mateo Sokač, Tafadzwa L Chaunzwa, Simon Bernatz, Ahmed Hosny, Raymond H Mak, Nicolai J Birkbak, et al. Foun- dation model for cancer imaging biomarkers.Nature machine intelligence, 6(3):354–367, 2024

    2024

  17. [17]

    Towards generalist foundation model for radiology by leveraging web-scale 2d&3d medical data.Nature Communications, 16(1):7866, 2025

    Chaoyi Wu, Xiaoman Zhang, Ya Zhang, Hui Hui, Yanfeng Wang, and Weidi Xie. Towards generalist foundation model for radiology by leveraging web-scale 2d&3d medical data.Nature Communications, 16(1):7866, 2025

    2025

  18. [18]

    An mri–pathology foundation model for noninvasive diagnosis and grading of prostate cancer.Nature Cancer, 6(10):1621–1637, 2025

    Lizhi Shao, Chao Liang, Ye Yan, Haibin Zhu, Xiaoming Jiang, Meiling Bao, Pan Zang, Xiazi Huang, Hongyu Zhou, Pei Nie, et al. An mri–pathology foundation model for noninvasive diagnosis and grading of prostate cancer.Nature Cancer, 6(10):1621–1637, 2025

    2025

  19. [19]

    A foundation model for lesion 17 segmentation on brain mri with mixture of modality experts.IEEE Transactions on Medical Imaging, 2025

    Xinru Zhang, Ni Ou, Berke Doga Basaran, Marco Visentin, Mengyun Qiao, Renyang Gu, Paul M Matthews, Yaou Liu, Chuyang Ye, and Wenjia Bai. A foundation model for lesion 17 segmentation on brain mri with mixture of modality experts.IEEE Transactions on Medical Imaging, 2025

    2025

  20. [20]

    Millennium pathways for tractography: 40 grand challenges to shape the future of tractography.ArXiv, pages arXiv–2509, 2025

    Maxime Descoteaux, Kurt G Schilling, Dogu Baran Aydogan, Christian Beaulieu, Elena Borra, Maxime Chamberland, Alessandro Daducci, Alberto De Luca, Flavio Dell’Acqua, Jessica Dubois, et al. Millennium pathways for tractography: 40 grand challenges to shape the future of tractography.ArXiv, pages arXiv–2509, 2025

    2025

  21. [21]

    An anatomically curated fiber clustering white matter atlas for consistent white matter tract parcellation across the lifespan.Neuroimage, 179:429–447, 2018

    Fan Zhang, Ye Wu, Isaiah Norton, Laura Rigolo, Yogesh Rathi, Nikos Makris, and Lauren J O’Donnell. An anatomically curated fiber clustering white matter atlas for consistent white matter tract parcellation across the lifespan.Neuroimage, 179:429–447, 2018

    2018

  22. [22]

    Deep white matter analysis (deepwma): Fast and consistent tractography segmentation.Med

    Fan Zhang, Suheyla Cetin Karayumak, Nico Hoffmann, Yogesh Rathi, Alexandra J Golby, and Lauren J O’Donnell. Deep white matter analysis (deepwma): Fast and consistent tractography segmentation.Med. Image Anal., 65:101761, 2020

    2020

  23. [23]

    Tractcloud: Registration-free tractography parcellation with a novel local-global streamline point cloud representation

    Tengfei Xue, Yuqian Chen, Chaoyi Zhang, Alexandra J Golby, Nikos Makris, Yogesh Rathi, Weidong Cai, Fan Zhang, and Lauren J O’Donnell. Tractcloud: Registration-free tractography parcellation with a novel local-global streamline point cloud representation. InProc. Int. Conf. Med. Image Comput. Comput. Assist. Interv., pages 409–419, 2023

    2023

  24. [24]

    Haotian Xu, Ming Dong, Min-Hee Lee, Nolan O’Hara, Eishi Asano, and Jeong-Won Jeong. Objective detection of eloquent axonal pathways to minimize postoperative deficits in pediatric epilepsy surgery using diffusion tractography and convolutional neural networks.IEEE Trans. Med. Imaging, 38(8):1910–1922, 2019

    1910

  25. [25]

    Dynamic graph cnn for learning on point clouds.ACM Trans

    Yue Wang, Yongbin Sun, Ziwei Liu, Sanjay E Sarma, Michael M Bronstein, and Justin M Solomon. Dynamic graph cnn for learning on point clouds.ACM Trans. Graph., 38(5):1–12, 2019

    2019

  26. [26]

    Rapidparc: A global-context transformer for parallel, accurate, and lesion- robust tractogram parcellation.Imaging Neuroscience, 4:IMAG–a, 2026

    Justus Bisten, Valentin von Bornhaupt, Johannes Grün, Tobias Bauer, Theodor Rüber, and Thomas Schultz. Rapidparc: A global-context transformer for parallel, accurate, and lesion- robust tractogram parcellation.Imaging Neuroscience, 4:IMAG–a, 2026

    2026

  27. [27]

    Prediction of individual subject’s age across the human lifespan using diffusion tensor imaging: a machine learning approach

    Benson Mwangi, Khader M Hasan, and Jair C Soares. Prediction of individual subject’s age across the human lifespan using diffusion tensor imaging: a machine learning approach. Neuroimage, 75:58–67, 2013

    2013

  28. [28]

    Modelandpredictageandsex inhealthysubjectsusing brain white matterfeatures: a deep learning approach

    Hao He, Fan Zhang, Steve Pieper, Nikos Makris, Yogesh Rathi, William Wells, and Lauren J O’Donnell. Modelandpredictageandsex inhealthysubjectsusing brain white matterfeatures: a deep learning approach. In2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI), pages 1–5. IEEE, 2022

    2022

  29. [29]

    Tractgraphcnn: anatomically informed graph cnn for classification using diffusion mri tractography

    Yuqian Chen, Fan Zhang, Leo R Zekelman, Tengfei Xue, Chaoyi Zhang, Yang Song, Nikos Makris, Yogesh Rathi, Weidong Cai, and Lauren J O’Donnell. Tractgraphcnn: anatomically informed graph cnn for classification using diffusion mri tractography. InInternational Sym- posium on Biomedical Imaging, pages 1–5. IEEE, 2023

    2023

  30. [30]

    Tractgraphformer: Anatomically informed hybrid graph cnn-transformer network for interpretable sex and age prediction from diffusion mri tractography.Med

    Yuqian Chen, Fan Zhang, Meng Wang, Leo R Zekelman, Suheyla Cetin-Karayumak, Tengfei Xue, Chaoyi Zhang, Yang Song, Jarrett Rushmore, Nikos Makris, et al. Tractgraphformer: Anatomically informed hybrid graph cnn-transformer network for interpretable sex and age prediction from diffusion mri tractography.Med. Image Anal., 101:103476, 2025. 18

    2025

  31. [31]

    Global fiber reconstruction becomes practical.Neuroimage, 54(2):955– 962, 2011

    Marco Reisert, Irina Mader, Constantin Anastasopoulos, Matthias Weigel, Susanne Schnell, and Valerij Kiselev. Global fiber reconstruction becomes practical.Neuroimage, 54(2):955– 962, 2011

    2011

  32. [32]

    Pointnet: Deep learning on point sets for 3d classification and segmentation

    Charles R Qi, Hao Su, Kaichun Mo, and Leonidas J Guibas. Pointnet: Deep learning on point sets for 3d classification and segmentation. InProc. IEEE Conf. Comput. Vis. Pattern Recognit., pages 652–660, 2017

    2017

  33. [33]

    Predicting age across hu- man lifespan based on structural connectivity from diffusion tensor imaging

    Cheol E Han, Luis R Peraza, John-Paul Taylor, and Marcus Kaiser. Predicting age across hu- man lifespan based on structural connectivity from diffusion tensor imaging. InIEEE BioCAS, pages 137–140, 2014

    2014

  34. [34]

    Predictability of intelligence and age from structural connec- tomes.Plos one, 19(4):e0301599, 2024

    Sebastian J Kopetzky, Yong Li, Marcus Kaiser, Markus Butz-Ostendorf, and Alzheimer’s Dis- ease Neuroimaging Initiative. Predictability of intelligence and age from structural connec- tomes.Plos one, 19(4):e0301599, 2024

    2024

  35. [35]

    brain-age

    Guruprasath Gurusamy, Varsha Sreenivasan, Ranjini Garani Ramesh, Simran Purokayastha, Mahendra Javali, Naren P Rao, and Devarajan Sridharan. Diffusion mri-based structural connectivity robustly predicts “brain-age”. InNeurIPS Workshop, 2020

    2020

  36. [36]

    Supervised contrastive learning.Advances in neural information processing systems, 33:18661–18673, 2020

    Prannay Khosla, Piotr Teterwak, Chen Wang, Aaron Sarna, Yonglong Tian, Phillip Isola, Aaron Maschinot, Ce Liu, and Dilip Krishnan. Supervised contrastive learning.Advances in neural information processing systems, 33:18661–18673, 2020

    2020

  37. [37]

    Self-supervised learning in medicine and healthcare.Nature Biomedical Engineering, 6(12):1346–1352, 2022

    Rayan Krishnan, Pranav Rajpurkar, and Eric J Topol. Self-supervised learning in medicine and healthcare.Nature Biomedical Engineering, 6(12):1346–1352, 2022

    2022

  38. [38]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, and Ross Girshick. Masked autoencoders are scalable vision learners. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 16000–16009, 2022

    2022

  39. [39]

    Predicting clinical outcome of stroke patients with tractographic feature

    Po-Yu Kao, Jeffereson W Chen, and BS Manjunath. Predicting clinical outcome of stroke patients with tractographic feature. InInternational MICCAI Brainlesion Workshop, pages 32–43. Springer, 2019

    2019

  40. [40]

    Precision medicine in stroke: towards personalized outcome predictions using artificial intelligence.Brain, 145(2):457–475, 2022

    Anna K Bonkhoff and Christian Grefkes. Precision medicine in stroke: towards personalized outcome predictions using artificial intelligence.Brain, 145(2):457–475, 2022

    2022

  41. [41]

    Tractography- based modeling explains treatment outcomes in patients undergoing deep brain stimulation for obsessive-compulsive disorder.Biological psychiatry, 96(2):95–100, 2024

    Ron Gadot, Ningfei Li, Ben Shofty, Michelle Avendano-Ortega, Sarah McKay, Kelly R Bijanki, Meghan E Robinson, Garrett Banks, Nicole Provenza, Eric A Storch, et al. Tractography- based modeling explains treatment outcomes in patients undergoing deep brain stimulation for obsessive-compulsive disorder.Biological psychiatry, 96(2):95–100, 2024

    2024

  42. [42]

    The wu-minn human connectome project: an overview.Neuroimage, 80:62–79, 2013

    David C Van Essen, Stephen M Smith, Deanna M Barch, Timothy EJ Behrens, Essa Yacoub, Kamil Ugurbil, Wu-Minn HCP Consortium, et al. The wu-minn human connectome project: an overview.Neuroimage, 80:62–79, 2013

    2013

  43. [43]

    The autism brain imaging data exchange: towards a large-scale evaluation of the intrinsic brain architecture in autism.Molecular psychiatry, 19(6):659–667, 2014

    Adriana Di Martino, Chao-Gan Yan, Qingyang Li, Erin Denio, Francisco X Castellanos, Kaat Alaerts, Jeffrey S Anderson, Michal Assaf, Susan Y Bookheimer, Mirella Dapretto, et al. The autism brain imaging data exchange: towards a large-scale evaluation of the intrinsic brain architecture in autism.Molecular psychiatry, 19(6):659–667, 2014. 19

    2014

  44. [44]

    The alzheimer’s disease neuroimaging initiative.Neuroimaging Clinics, 15(4):869–877, 2005

    Susanne G Mueller, Michael W Weiner, Leon J Thal, Ronald C Petersen, Clifford Jack, William Jagust, John Q Trojanowski, Arthur W Toga, and Laurel Beckett. The alzheimer’s disease neuroimaging initiative.Neuroimaging Clinics, 15(4):869–877, 2005

    2005

  45. [45]

    A phenome-wide examination of neural and cognitive function.Scientific data, 3(1):160110, 2016

    Russell A Poldrack, Eliza Congdon, William Triplett, Krzysztof J Gorgolewski, Katherine H Karlsgodt, Jonathan A Mumford, Farrah W Sabb, Nelson B Freimer, Edythe D London, Ty- rone D Cannon, et al. A phenome-wide examination of neural and cognitive function.Scientific data, 3(1):160110, 2016

    2016

  46. [46]

    The parkinson progression marker initiative (ppmi).Progress in neurobiology, 95(4):629–635, 2011

    Kenneth Marek, Danna Jennings, Shirley Lasch, Andrew Siderowf, Caroline Tanner, Tanya Simuni, Chris Coffey, Karl Kieburtz, Emily Flagg, Sohini Chowdhury, et al. The parkinson progression marker initiative (ppmi).Progress in neurobiology, 95(4):629–635, 2011

    2011

  47. [47]

    Scikit- learn: Machine learning in python.Journal of machine learning research, 12:2825–2830, 2011

    Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, et al. Scikit- learn: Machine learning in python.Journal of machine learning research, 12:2825–2830, 2011. 20

    2011

This paper was first reviewed by grok-4.3 on June 28, 2026.