Pith. sign in

REVIEW 5 major objections 5 minor 57 references

DeepMultiConnectome: Deep Multi-Task Prediction of Structural Connectomes Directly from Diffusion MRI Tractography

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read DeepMultiConnectome predicts structural connectomes directly from diffusion MRI tractography, matching the traditional pipeline at r≈0.99 while cutting the per-subject parcellation bottleneck.

desk verdict DeepMultiConnectome is a genuinely useful engineering contribution for fast multi-atlas connectome generation, but its headline correlation numbers are less decisive than they look—and the authors themselves admit that risk. read the letter →

arxiv 2505.22685 v2 pith:F5PHF4LO submitted 2025-05-27 eess.IV cs.AIcs.CV

Marcus J. Vroemen , Yuqian Chen , Yui Lo , Tengfei Xue , Weidong Cai , Fan Zhang , Josien P.W. Pluim , Lauren J. O'Donnell This is my paper · ORCID
classification eess.IVcs.AIcs.CV
keywords structuralconnectomediffusionMRItractographypointcloudneuralnetworkmulti-tasklearningpredictionparcellation-freebrainmeasurestest-retestreproducibility
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper sets out to prove that a structural connectome can be predicted straight from a whole-brain tractogram, skipping the gray-matter parcellation step that makes the traditional pipeline slow. The proposed DeepMultiConnectome classifies each streamline into the pair of regions it connects, using a shared representation to predict connectomes for an 84-region and a 164-region parcellation at the same time. Across 200 unseen brains, whole connectomes are generated in about 40 seconds and correlate with traditional connectomes at $r=0.992\pm0.003$ (84 regions) and $r=0.986\pm0.004$ (164 regions). Test-retest reproducibility and downstream age/cognition predictions are statistically comparable to the traditional method, which is what would make the fast surrogate useful.

What carries the argument

The load-bearing object is a point-cloud classifier in which every streamline is resampled to 15 evenly spaced 3D coordinates and mapped to a class label standing for an unordered pair of gray-matter regions, plus an unknown class for unassignable streamlines. A multi-task output layer shares the learned streamline representation and predicts the 84-region and 164-region pair labels in parallel; tallying the predicted labels over all streamlines of a tractogram yields the two connectome matrices.

What would settle it

Compute Pearson correlation between predicted and traditional connectomes separately for edges grouped by traditional streamline count, such as the lowest quintile against the highest quintile. If low-count edges show near-zero correlation while overall $r\approx0.99$ stays high, the headline similarity is carried by high-count edges and per-edge fidelity is not established.

Watch

Extended reading notes

Core claim

The central claim is that streamline geometry alone is enough to recover the gray-matter endpoint pair of each fiber, so the entire connectome can be assembled by classifying streamlines rather than by parcellating anatomy. The model emits two connectomes per tractogram in a single forward pass, and the paper reports that these predicted matrices are nearly indistinguishable from traditional ones by Pearson correlation, preserve standard network measures, reproduce across scan sessions as well as the traditional pipeline, and carry comparable subject-specific signal for age and cognition prediction.

Load-bearing premise

That 15 evenly spaced sample points along a streamline carry enough information to identify which pair of gray-matter regions it connects, so a learned geometry-to-label map can replace anatomical parcellation.

Editorial extensions

If this is right

  • Whole-brain connectomes can be produced for large cohorts without running any anatomical parcellation at inference time.
  • A single trained model serves multiple parcellation schemes simultaneously, so atlas choice no longer multiplies compute.
  • Predicted connectomes preserve enough network-level information to support studies of integration, segregation, and resilience comparable to traditional matrices.
  • Subject-level reproducibility is maintained, so longitudinal or test-retest designs can use the fast surrogate without obvious loss.

Reading between the lines

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

  • The shared representation should extend to additional atlases by adding output heads, since all schemes describe the same streamlines; a direct test is whether a third parcellation can be learned without retraining the backbone.
  • The wide gap between accuracy (~81% and ~70%) and macro F1 (~50% and ~34%) implies that rare, low-count edges are often wrong even though the matrix correlation is high; conclusions about long-range or sparse connections should not be drawn from these matrices until per-edge fidelity is shown.
  • The evaluation ceiling is the conventional nearest-endpoint labeling, so any systematic bias in that definition of connectivity is inherited exactly; the method cannot be more anatomically correct than its training target.
  • A practical extension would be to use the classifier's confidence as a per-streamline quality filter within the same 40-second pass, potentially improving downstream connectomes without a separate filtering step.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The paper proposes DeepMultiConnectome, a point-cloud neural network that classifies tractography streamlines directly into gray-matter region pairs for two parcellation schemes (84 and 164 ROIs) simultaneously, thereby bypassing anatomical parcellation. Trained on 1,000 HCP-YA subjects with labels derived from MRtrix's tck2connectome, the model predicts connectomes from approximately 2.93 million streamlines per subject in about 40 seconds. On 200 held-out test subjects, predicted connectomes achieve Pearson correlations of r=0.992 (84 ROI) and r=0.986 (164 ROI) with traditional connectomes, largely preserve six network measures, show test-retest reproducibility statistically indistinguishable from the traditional pipeline, and yield comparable age/cognition prediction performance in a downstream 1D-CNN. The authors conclude that DeepMultiConnectome provides a scalable and fast surrogate for the conventional parcellation-and-labeling connectome pipeline.

Significance. If the subject-specific predictive value is firmly established, the method is valuable: it reduces per-subject connectome generation from hours to about 40 seconds, removes the FreeSurfer parcellation bottleneck, and supports multiple parcellation schemes in a single model. The evaluation is out-of-sample on 200 unseen subjects and includes several complementary checks—intrasubject versus intersubject similarity, network-measure correlations, test-retest parity, and a downstream phenotype prediction experiment—which are appropriate and largely behave as expected for a successful surrogate. The paper is also transparent about its limitations, including the explicit concession in the Discussion that its evaluation methods may overestimate true prediction quality. However, the central quantitative claim currently rests on Pearson correlations over sparse connectome matrices, which can be inflated by stereotypy and zero-zero agreement; the absence of a group-average baseline and edge-wise analyses leaves the subject-specificity of the predictions not fully established.

major comments (5)
  1. [Section 3.2, Figure 4, Table 3] The headline r=0.992/0.986 is never compared against the simplest null model: assigning the mean training connectome to every test subject. Because structural connectomes are highly stereotyped across healthy young adults and the matrices are sparse, Pearson correlation over upper-triangular elements can be high even for a group-average predictor. The paper reports intrasubject versus intersubject similarity only as distributions in Figure 4 without giving the actual intersubject correlation values. Please report those values and add a group-average baseline to the same evaluation. This is load-bearing for the claim that the predictions are subject-specific rather than reflections of population-level connectivity.
  2. [Section 3.1, Tables 1 and 2; Section 3.2] The macro F1 scores of 50.48% (84 ROI) and 34.36% (164 ROI) indicate that the majority of streamline classes are misclassified at the per-class level. The paper hypothesizes that these errors are localized to low-impact edges or that the evaluation is insensitive, but this is not demonstrated. Please provide edge-wise analyses: for example, prediction error stratified by true edge count, Pearson correlation computed only over nonzero edges, and confusion statistics for low-count and long-range edges. Without such analyses, the high connectome-level r could be carried by a few high-count edges and by agreement on zero-zero pairs, and the stated macro F1 makes this a concrete risk.
  3. [Section 3.3, Table 4] The test-retest parity between DeepMultiConnectome and the traditional method is presented as evidence of reproducibility, but both methods are evaluated against the same tck2connectome labeling procedure, and a group-average predictor would also show high test-retest correlation because the same group mean would be assigned to both sessions. Please add a group-average baseline to Table 4 and report the actual intersubject test-retest or between-subject similarity values. This would allow the parity to be interpreted as subject-specific reliability rather than as a consequence of stereotypy.
  4. [Section 2.2.2 and Section 3.2] The training labels and the evaluation target are both generated by tck2connectome's nearest-endpoint assignment from the same parcellation. The paper should explicitly state that the reported r measures agreement with this specific automated pipeline and not with an independent anatomical ground truth, and it should discuss how this ceiling affects the claim that the model can replace the traditional pipeline. The Discussion partially acknowledges a ceiling, but the statement belongs in the methods and evaluation sections where the metric is introduced and interpreted.
  5. [Section 3.4, Table 5] The downstream age and cognition prediction experiment is a valuable usability check, but it should include a null baseline to interpret the reported metrics. For age, the MAE of about 3.5 years should be compared with the MAE of predicting the mean age; for cognitive scores, a permutation test or a zero-correlation null would clarify whether the reported r values exceed chance. With n=200 and a high-dimensional 1D-CNN input, overfitting is a concern even with cross-validation, and the fold-wise standard deviations are relatively large. Adding these baselines would strengthen the conclusion that predicted connectomes retain subject-specific phenotypic information.
minor comments (5)
  1. [Section 2.4] The text states that training used '8 million streamlines from the training dataset, including 10,000 streamlines per training subject'; with 700 training subjects, 10,000 per subject yields 7 million, not 8 million. Please clarify the discrepancy or correct the number.
  2. [Table 5 and Section 2.1] The abbreviation 'TF AT' in Table 5 appears to be a typo for 'TFAT' (Flanker Attention and Inhibitory Control Test); the abbreviation list in Section 2.1 should be made consistent.
  3. [Section 2.2.1] The sentence 'The interface was then co-registered to dMRI space using masks provided by the HCP-YA dataset' is unclear; it seems to refer to using HCP-provided transforms or segmentations rather than masks. Please rephrase for clarity.
  4. [Figure 4 caption] The caption states that a Wilcoxon signed-rank test was used to compare intrasubject and intersubject similarity, but these two samples are not naturally paired (each subject has one intrasubject value and many intersubject values). Please specify exactly how the paired test was constructed.
  5. [Section 3.2] The reported LERM distances of 7.25±1.35 and 17.03±1.29 should be explicitly identified as intrasubject distances; the text currently introduces them without stating whether they are within-subject or also include a comparison to intersubject values.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the connectome predictions are evaluated out-of-sample on unseen subjects, and the self-citations are architectural provenance rather than load-bearing proof.

full rationale

The central claim is that DeepMultiConnectome predicts structural connectomes that closely match traditional tck2connectome-generated connectomes. This is a supervised surrogate-model claim, not a circular one: the training labels and the evaluation target are both produced by the same tck2connectome procedure (Section 2.2.2), but the model is evaluated on 200 test subjects whose tractograms and labels were not used for training, and the test-retest and phenotype-prediction experiments provide independent checks. A model can be trained to mimic a reference pipeline and then genuinely predict that pipeline's output on new data; the shared reference standard defines the task's ceiling, not a circular derivation. The self-citations to TractCloud and related point-cloud work (e.g., [52], [53]) are architectural building blocks, and the paper explicitly ablates the local-global embedding, ultimately selecting the simpler single-streamline PointNet (Table 1), so the cited prior work is not what forces the reported r values. There is no fitted parameter renamed as a prediction, no imported uniqueness theorem, and no definitional equivalence between an input and an output. The Discussion's admission that 'our evaluation methods may not be sufficiently sensitive, potentially overestimating the true quality of the predictions' is a valid validity concern about baseline sensitivity and metric choice, but it is not a circularity step. Therefore the paper's derivation chain is self-contained and no circularity is identified.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

No new physical, biological, or mathematical entities are postulated. The network is a standard trained model, and the two parcellation schemes are pre-existing FreeSurfer atlases, so no invented entities were introduced. The free parameters and axioms above cover the fitted and assumed content the central claim rests on.

free parameters (6)
  • Neural network weights (PointNet backbone) = streamline accuracy 81.23% (84 ROI), 70.15% (164 ROI) on validation
    The connectome prediction function is entirely the trained network; weights were fit to 8M labeled streamlines from 700 training subjects (Section 2.4). All headline results depend on this fitted object.
  • Learning rate = 0.001 (selected from 0.01, 0.001, 0.0001)
    Chosen by a small grid search on the 100-subject validation set (Section 2.4).
  • Batch size = 1,024 (selected from 512, 1,024, 2,048)
    Chosen by grid search on validation (Section 2.4).
  • Streamline resample points = 15 equally spaced 3D coordinates
    Hand-fixed representation following prior point-cloud tractography work [56]; the model sees only these points, so the central geometric assumption depends on this choice (Section 2.3.2).
  • Training streamlines per subject = 10,000 per subject, 8M total, from 700 subjects
    Fixed following prior work [52,53] (Section 2.4); undersamples rare node-pair classes, plausibly explaining the low macro F1 (50.5% and 34.4%).
  • Class label scheme (lexicographic node pairs plus unknown class) = 3,571 classes (84 ROI); 13,631 classes (164 ROI)
    Design choice in Section 2.3.1 that defines the classification problem; the additional 'unknown' class absorbs unassignable streamlines and affects reported accuracy.
assumptions (5)
  • domain assumption tck2connectome's nearest-endpoint region assignment is an accurate ground truth for streamline connectivity.
    Section 2.2.2: this procedure generates both the training labels and the evaluation target; any systematic error in endpoint-to-ROI assignment is an unreachable ceiling on predicted connectomes because the model can only imitate this labeling.
  • domain assumption A streamline resampled to 15 3D points retains enough geometric information to determine its connected region pair.
    Section 2.3.2: the network never sees the T1 anatomy or any information beyond the 15 points; the Discussion acknowledges gyral variability causes adjacent-region misclassifications.
  • domain assumption The HCP-YA 1,000-subject cohort is representative of the populations where the model will be used.
    Section 2.1: healthy adults aged 22-35 scanned on one customized scanner; the paper lists lifespan and cross-acquisition generalization as future work in Section 4.
  • domain assumption Whole-brain probabilistic tractography (MRtrix3 iFOD2 with ACT) yields streamlines whose endpoint pairs reflect true structural connections.
    Section 2.2.1: both traditional and predicted connectomes inherit tractography's known false-positive and premature-termination limitations; the paper evaluates fidelity to the pipeline, not biological truth.
  • standard math PointNet without T-Net, as configured in the cited TractCloud/DeepWMA work, is an adequate classifier for this task.
    Section 2.3.2: the architecture is borrowed from prior published work [52,53] and assumed applicable; no architectural derivation is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DeepMultiConnectome: Deep Multi-Task Prediction of Structural Connectomes Directly from Diffusion MRI Tractography." pith.science (2026). https://pith.science/paper/F5PHF4LO

@misc{pith2026250522685,
  author       = {Pith},
  title        = {Pith review of: DeepMultiConnectome: Deep Multi-Task Prediction of Structural Connectomes Directly from Diffusion MRI Tractography},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F5PHF4LO}},
  note         = {Machine review of arXiv:2505.22685}
}
abstract

Diffusion MRI (dMRI) tractography enables in vivo mapping of brain structural connections, but traditional connectome generation is time-consuming and requires gray matter parcellation, posing challenges for large-scale studies. We introduce DeepMultiConnectome, a deep-learning model that predicts structural connectomes directly from tractography, bypassing the need for gray matter parcellation while supporting multiple parcellation schemes. Using a point-cloud-based neural network with multi-task learning, the model classifies streamlines according to their connected regions across two parcellation schemes, sharing a learned representation. We train and validate DeepMultiConnectome on tractography from the Human Connectome Project Young Adult dataset ($n = 1000$), labeled with an 84 and 164 region gray matter parcellation scheme. DeepMultiConnectome predicts multiple structural connectomes from a whole-brain tractogram containing 3 million streamlines in approximately 40 seconds. DeepMultiConnectome is evaluated by comparing predicted connectomes with traditional connectomes generated using the conventional method of labeling streamlines using a gray matter parcellation. The predicted connectomes are highly correlated with traditionally generated connectomes ($r = 0.992$ for an 84-region scheme; $r = 0.986$ for a 164-region scheme) and largely preserve network properties. A test-retest analysis of DeepMultiConnectome demonstrates reproducibility comparable to traditionally generated connectomes. The predicted connectomes perform similarly to traditionally generated connectomes in predicting age and cognitive function. Overall, DeepMultiConnectome provides a scalable, fast model for generating subject-specific connectomes across multiple parcellation schemes.

Figures

Figures reproduced from arXiv: 2505.22685 by the authors.

Figure 1
Figure 1. Structural connectome construction using the traditional and proposed method. The traditional method requires an anatomical parcellation of the brain, which can be computationally expensive. Our method avoids this by using a deep learning network to predict the anatomical regions that streamlines connect. Advances in deep learning have demonstrated the potential for fast and consistent segmentation of the brain’s wh… view at source ↗
Figure 2
Figure 2. Diagram of the data preparation pipeline, showing how the MRI data is used to generate whole￾brain tractograms and structural connectomes. The two different gray matter parcellations are used to give two label sets to each streamline and generate two structural connectomes. The streamlines (S) and assigned labels (L1 and L2) form the training dataset. 2.2.1 Tractography Whole-brain tractograms were generated using p… view at source ↗
Figure 3
Figure 3. Diagram of the inference procedure of DeepMultiConnectome. One single streamline is used as input for the model, which predicts a class that is associated with the node pair it connects. Predictions of all the streamlines in a whole-brain tractogram are summarized in a connectome matrix. This figure shows the simultaneous prediction of two sets of node pairs according to two different parcellation schemes, using the… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Similarity of predicted and traditional connectomes. Violin plots show the distributions of intra￾subject (blue) and intersubject (red) correlation coefficients and log-Euclidean Riemannian metric (LERM) distances (n = 200). Connectomes of both parcellation schemes are…
Figure 5
Figure 5. Figure 5: Traditional and DeepMultiConnectome predicted connectomes based on the 84 and 164 ROI parcellation schemes of an example test subject. The difference map was created by subtracting the predicted from the traditional connectome, and it shows the over- or underclassifica…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 57 canonical work pages

  1. [1]

    Arsigny, V., Fillard, P., Pennec, X., and Ayache, N. (2006). Log-euclidean metrics for fast and simple calculus on diffusion tensors.Magn. Reson. Med., 56(2):411–421

    work page 2006

  2. [2]

    I., Makropoulos, A., Robinson, E

    Arslan, S., Ktena, S. I., Makropoulos, A., Robinson, E. C., Rueckert, D., and Parisot, S. (2018). Hu- man brain mapping: A systematic comparison of parcellation methods for the human cerebral cortex. Neuroimage, 170:5–30

    work page 2018

  3. [3]

    Astolfi, P., Verhagen, R., Petit, L., Olivetti, E., Masci, J., Boscaini, D., and Avesani, P. (2020). Trac- togram filtering of anatomically non-plausible fibers with geometric deep learning. InMedical Image Computing and Computer Assisted Intervention – MICCAI 2020, Lecture notes in computer science, pages 291–301. Springer International Publishing, Cham

    work page 2020

  4. [4]

    Baecker, L., Garcia-Dias, R., Vieira, S., Scarpazza, C., and Mechelli, A. (2021). Machine learning for brain age prediction: Introduction to methods and clinical applications.EBioMedicine, 72(103600):103600

    work page 2021

  5. [5]

    Bassett, D. S. and Bullmore, E. T. (2009). Human brain networks in health and disease.Curr. Opin. Neurol., 22(4):340–347

    work page 2009

  6. [6]

    and Tavor, I

    Bernstein-Eliav, M. and Tavor, I. (2024). The prediction of brain activity from connectivity: Advances and applications.Neuroscientist, 30(3):367–377

    work page 2024

  7. [7]

    J., Deems, S., Furlani, T

    Boerner, T. J., Deems, S., Furlani, T. R., Knuth, S. L., and Towns, J. (2023). ACCESS: Advancing innovation: NSF’s advanced cyberinfrastructure coordination ecosystem: Services & support. InPractice and Experience in Advanced Research Computing, New York, NY, USA. ACM

    work page 2023

  8. [8]

    Cao, M., Huang, H., Peng, Y., Dong, Q., and He, Y. (2016). Toward developmental connectomics of the human brain.Front. Neuroanat., 10:25

    work page 2016

Show all 57 references
  1. [9]

    Chen, Y., Zhang, C., Xue, T., Song, Y., Makris, N., Rathi, Y., Cai, W., Zhang, F., and O’Donnell, L. J. (2023a). Deep fiber clustering: Anatomically informed fiber clustering with self-supervised deep learning for fast and effective tractography parcellation.Neuroimage, 273:120086

    2023

  2. [10]

    R., Xue, T., Zhang, C., Song, Y., Makris, N., Rathi, Y., Cai, W., and O’Donnell, L

    Chen, Y., Zhang, F., Zekelman, L. R., Xue, T., Zhang, C., Song, Y., Makris, N., Rathi, Y., Cai, W., and O’Donnell, L. J. (2023b). Tractgraphcnn: Anatomically informed graph CNN for classification using diffusion MRI tractography. In2023 IEEE 20th International Symposium on Bio...

    2023

  3. [11]

    R., He, J., Song, Y., Makris, N., Rathi, Y., Golby, A

    Chen, Y., Zhang, F., Zhang, C., Xue, T., Zekelman, L. R., He, J., Song, Y., Makris, N., Rathi, Y., Golby, A. J., Cai, W., and O’Donnell, L. J. (2022). White matter tracts are point clouds: Neuropsychological score prediction and critical region localization via geometric deep ...

    2022

  4. [12]

    S., S´ egonne, F., Fischl, B., Quinn, B

    Desikan, R. S., S´ egonne, F., Fischl, B., Quinn, B. T., Dickerson, B. C., Blacker, D., Buckner, R. L., Dale, A. M., Maguire, R. P., Hyman, B. T., Albert, M. S., and Killiany, R. J. (2006). An automated labeling system for subdividing the human cerebral cortex on MRI scans int...

    2006

  5. [13]

    Destrieux, C., Fischl, B., Dale, A., and Halgren, E. (2010). Automatic parcellation of human cortical gyri and sulci using standard anatomical nomenclature.Neuroimage, 53(1):1–15. 13 DeepMultiConnectome: Predicting Structural Connectomes

    2010

  6. [14]

    Dhollander, T., Mito, R., Raffelt, D., and Connelly, A. (2019). Improved white matter response function estimation for 3-tissue constrained spherical deconvolution. InProc. Intl. Soc. Mag. Reson. Med, volume

    2019

  7. [15]

    Dhollander, T., Raffelt, D., and Connelly, A. (2016). Unsupervised 3-tissue response function estimation from single-shell or multi-shell diffusion MR data without a co-registered T1 image

    2016

  8. [16]

    Dhollander, T., Tabbara, R., Rosnarho-Tornstrand, J., Tournier, J.-D., Raffelt, D., and Connelly, A. (2021). Multi-tissue log-domain intensity and inhomogeneity normalisation for quantitative apparent fibre density. InProc. ISMRM, volume 29, page 2472. archive.ismrm.org

    2021

  9. [17]

    Fischl, B. (2012). FreeSurfer.Neuroimage, 62(2):774–781

    2012

  10. [18]

    Fishburn, P. C. (1974). Exceptional paper—lexicographic orders, utilities and decision rules: A survey. Manage. Sci., 20(11):1442–1471

    1974

  11. [19]

    C., Slotkin, J., Manly, J

    Gershon, R. C., Slotkin, J., Manly, J. J., Blitz, D. L., Beaumont, J. L., Schnipke, D., Wallner-Allen, K., Golinkoff, R. M., Gleason, J. B., Hirsh-Pasek, K., Adams, M. J., and Weintraub, S. (2013). IV. NIH toolbox cognition battery (CB): measuring language (vocabulary comprehe...

    2013

  12. [20]

    F., Sotiropoulos, S

    Glasser, M. F., Sotiropoulos, S. N., Wilson, J. A., Coalson, T. S., Fischl, B., Andersson, J. L., Xu, J., Jbabdi, S., Webster, M., Polimeni, J. R., Van Essen, D. C., Jenkinson, M., and WU-Minn HCP Consortium (2013). The minimal preprocessing pipelines for the human connectome ...

    2013

  13. [21]

    C., Grumbach, P., Goltermann, J., Winter, N

    Gruber, M., Mauritz, M., Meinert, S., Grotegerd, D., de Lange, S. C., Grumbach, P., Goltermann, J., Winter, N. R., Waltemate, L., Lemke, H., Thiel, K., Winter, A., Breuer, F., Borgers, T., Enneking, V., Klug, M., Brosch, K., Meller, T., Pfarr, J.-K., Ringwald, K. G., Stein, F....

    2023

  14. [22]

    Y., Fischer, J., Lowe, J

    Hancock, D. Y., Fischer, J., Lowe, J. M., Snapp-Childs, W., Pierce, M., Marru, S., Coulter, J. E., Vaughn, M., Beck, B., Merchant, N., Skidmore, E., and Jacobs, G. (2021). Jetstream2: Accelerating cloud computing via jetstream. InPractice and Experience in Advanced Research Co...

    2021

  15. [23]

    He, H., Zhang, F., Pieper, S., Makris, N., Rathi, Y., Wells, W., and O’Donnell, L. J. (2022). Model and predict age and sex in healthy subjects using brain white matter features: A deep learning approach. In 2022 IEEE 19th International Symposium on Biomedical Imaging (ISBI), ...

    2022

  16. [24]

    Henschel, L., Conjeti, S., Estrada, S., Diers, K., Fischl, B., and Reuter, M. (2020). FastSurfer - a fast and accurate deep learning based neuroimaging pipeline.Neuroimage, 219(117012):117012

    2020

  17. [25]

    Jeurissen, B., Tournier, J.-D., Dhollander, T., Connelly, A., and Sijbers, J. (2014). Multi-tissue con- strained spherical deconvolution for improved analysis of multi-shell diffusion MRI data.Neuroimage, 103:411–426

    2014

  18. [26]

    C., Velasco-Annis, C., Rollins, C., Jaimes, C., Gholipour, A., and Warfield, S

    Karimi, D., Calixto, C., Snoussi, H., Cortes-Albornoz, M. C., Velasco-Annis, C., Rollins, C., Jaimes, C., Gholipour, A., and Warfield, S. K. (2024). Detailed delineation of the fetal brain in diffusion MRI via multi-task learning.bioRxivorg

    2024

  19. [27]

    D., Ferrall, G., Patterson, J., Styner, B., and Prieto, M

    Lam, N., Belhomme, P. D., Ferrall, G., Patterson, J., Styner, B., and Prieto, M. (2018). TRAFIC: Fiber tract classification using deep learning.Proc. SPIE Int. Soc. Opt. Eng

    2018

  20. [28]

    Liu, F., Feng, J., Chen, G., Wu, Y., Hong, Y., Yap, P.-T., and Shen, D. (2019). DeepBundle: Fiber bundle parcellation with graph convolution neural networks.arXiv [eess.IV]. M.J. Vroemen et al. 14

    2019

  21. [29]

    Liu, W., Chen, Y., Ye, C., Makris, N., Rathi, Y., Cai, W., Zhang, F., and O’Donnell, L. J. (2023). Fiber tract shape measures inform prediction of non-imaging phenotypes.arXiv [cs.CV]

    2023

  22. [30]

    A., and Zalesky, A

    Liu, Y., Seguin, C., Mansour L., S., Tian, Y., Di Biase, M. A., and Zalesky, A. (2025). Deep generation of personalized connectomes based on individual attributes.bioRxiv

    2025

  23. [31]

    H., Zekelman, L., Zhang, F., Rushmore, J., Rathi, Y., Makris, N., Golby, A

    Lo, Y., Chen, Y., Liu, D., Legarreta, J. H., Zekelman, L., Zhang, F., Rushmore, J., Rathi, Y., Makris, N., Golby, A. J., Cai, W., and O’Donnell, L. J. (2024a). TractShapeNet: Efficient multi-shape learning with 3D tractography point clouds.arXiv [cs.CV]

    2024

  24. [32]

    J., Cai, W., and O’Donnell, L

    Lo, Y., Chen, Y., Liu, D., Liu, W., Zekelman, L., Rushmore, J., Zhang, F., Rathi, Y., Makris, N., Golby, A. J., Cai, W., and O’Donnell, L. J. (2025). The shape of the brain’s connections is predictive of cognitive performance: An explainable machine learning study.Hum. Brain M...

    2025

  25. [33]

    J., Cai, W., et al

    Lo, Y., Chen, Y., Liu, D., Liu, W., Zekelman, L., Zhang, F., Rathi, Y., Makris, N., Golby, A. J., Cai, W., et al. (2024b). Cross-domain fiber cluster shape analysis for language performance cognitive score prediction. InInternational Workshop on Computational Diffusion MRI, pa...

    2024

  26. [34]

    A., Parker, D., and Verma, R

    Osmanlıo˘ glu, Y., Alappatt, J. A., Parker, D., and Verma, R. (2020). Connectomic consistency: a systematic stability analysis of structural and functional connectivity.J. Neural Eng., 17(4):045004

    2020

  27. [35]

    Qi Charles, R., Su, H., Kaichun, M., and Guibas, L. J. (2017). PointNet: Deep learning on point sets for 3D classification and segmentation. In2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 77–85. IEEE

    2017

  28. [36]

    and Sporns, O

    Rubinov, M. and Sporns, O. (2010). Complex network measures of brain connectivity: uses and inter- pretations.Neuroimage, 52(3):1059–1069

    2010

  29. [37]

    Sarwar, T., Seguin, C., Ramamohanarao, K., and Zalesky, A. (2020). Towards deep learning for con- nectome mapping: A block decomposition framework.Neuroimage, 212(116654):116654

    2020

  30. [38]

    S., Scheinost, D., Rosenberg, M

    Shen, X., Finn, E. S., Scheinost, D., Rosenberg, M. D., Chun, M. M., Papademetris, X., and Constable, R. T. (2017). Using connectome-based predictive modeling to predict individual behavior from brain connectivity.Nat. Protoc., 12(3):506–518

    2017

  31. [39]

    E., Tournier, J.-D., Calamante, F., and Connelly, A

    Smith, R. E., Tournier, J.-D., Calamante, F., and Connelly, A. (2012). Anatomically-constrained tractography: improved diffusion MRI streamlines tractography through effective use of anatomical infor- mation.Neuroimage, 62(3):1924–1938

    2012

  32. [40]

    N., Jbabdi, S., Xu, J., Andersson, J

    Sotiropoulos, S. N., Jbabdi, S., Xu, J., Andersson, J. L., Moeller, S., Auerbach, E. J., Glasser, M. F., Hernandez, M., Sapiro, G., Jenkinson, M., Feinberg, D. A., Yacoub, E., Lenglet, C., Van Essen, D. C., Ugurbil, K., Behrens, T. E. J., and WU-Minn HCP Consortium (2013). Adv...

    2013

  33. [41]

    Soumya Kumari, L. K. and Sundarrajan, R. (2024). A review on brain age prediction models.Brain Res., 1823(148668):148668

    2024

  34. [42]

    Sporns, O., Tononi, G., and K¨ otter, R. (2005). The human connectome: A structural description of the human brain.PLoS Comput. Biol., 1(4):e42

    2005

  35. [43]

    W., Clark, K

    Toga, A. W., Clark, K. A., Thompson, P. M., Shattuck, D. W., and Van Horn, J. D. (2012). Mapping the human connectome.Neurosurgery, 71(1):1–5

    2012

  36. [44]

    D., Calamante, F., Connelly, A., and Others (2010)

    Tournier, J. D., Calamante, F., Connelly, A., and Others (2010). Improved probabilistic streamlines tractography by 2nd order integration over fibre orientation distributions. InProceedings of the interna- tional society for magnetic resonance in medicine, volume 1670. archive...

    2010

  37. [45]

    G., and Connelly, A

    Tournier, J.-D., Calamante, F., Gadian, D. G., and Connelly, A. (2004). Direct estimation of the fiber orientation density function from diffusion-weighted MRI data using spherical deconvolution.Neuroimage, 23(3):1176–1185. 15 DeepMultiConnectome: Predicting Structural Connectomes

    2004

  38. [46]

    Tournier, J.-D., Smith, R., Raffelt, D., Tabbara, R., Dhollander, T., Pietsch, M., Christiaens, D., Jeurissen, B., Yeh, C.-H., and Connelly, A. (2019). MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation.Neuroimage

    2019

  39. [47]

    C., Smith, S

    Van Essen, D. C., Smith, S. M., Barch, D. M., Behrens, T. E. J., Yacoub, E., Ugurbil, K., and WU-Minn HCP Consortium (2013). The WU-minn human connectome project: an overview.Neuroimage, 80:62–79

    2013

  40. [48]

    J., Mansour, S., Kerrebijn, I., Medland, S

    Wainberg, M., Forde, N. J., Mansour, S., Kerrebijn, I., Medland, S. E., Hawco, C., and Tripathy, S. J. (2024). Genetic architecture of the structural connectome.Nat. Commun., 15(1):1962

    2024

  41. [49]

    E., Bronstein, M

    Wang, Y., Sun, Y., Liu, Z., Sarma, S. E., Bronstein, M. M., and Solomon, J. M. (2019). Dynamic graph CNN for learning on point clouds.ACM Trans. Graph., 38(5):1–12

    2019

  42. [50]

    Wasserthal, J., Neher, P., and Maier-Hein, K. H. (2018). TractSeg - fast and accurate white matter tract segmentation.Neuroimage, 183:239–253

    2018

  43. [51]

    S., Heaton, R

    Weintraub, S., Dikmen, S. S., Heaton, R. K., Tulsky, D. S., Zelazo, P. D., Bauer, P. J., Carlozzi, N. E., Slotkin, J., Blitz, D., Wallner-Allen, K., Fox, N. A., Beaumont, J. L., Mungas, D., Nowinski, C. J., Richler, J., Deocampo, J. A., Anderson, J. E., Manly, J. J., Borosh, B...

    2013

  44. [52]

    J., Makris, N., Rathi, Y., Cai, W., Zhang, F., and O’Donnell, L

    Xue, T., Chen, Y., Zhang, C., Golby, A. J., Makris, N., Rathi, Y., Cai, W., Zhang, F., and O’Donnell, L. J. (2023a). TractCloud: Registration-free tractography parcellation with a novel local-global streamline point cloud representation. InLecture Notes in Computer Science, Le...

    2023

  45. [53]

    J., Makris, N., Rathi, Y., Cai, W., and O’Donnell, L

    Xue, T., Zhang, F., Zhang, C., Chen, Y., Song, Y., Golby, A. J., Makris, N., Rathi, Y., Cai, W., and O’Donnell, L. J. (2023b). Superficial white matter analysis: An efficient point-cloud-based deep learning framework with supervised contrastive learning for consistent tractogr...

    2023

  46. [54]

    Zalesky, A., Sarwar, T., Tian, Y., Liu, Y., Yeo, B. T. T., and Ramamohanarao, K. (2024). Predicting an individual’s functional connectivity from their structural connectome: Evaluation of evidence, recom- mendations, and future prospects.Netw. Neurosci., 8(4):1291–1309

    2024

  47. [55]

    J., and O’Donnell, L

    Zhang, F., Cetin Karayumak, S., Hoffmann, N., Rathi, Y., Golby, A. J., and O’Donnell, L. J. (2020). Deep white matter analysis (DeepWMA): Fast and consistent tractography segmentation.Med. Image Anal., 65(101761):101761

    2020

  48. [56]

    Zhang, F., Wu, Y., Norton, I., Rigolo, L., Rathi, Y., Makris, N., and O’Donnell, L. J. (2018). An anatomically curated fiber clustering white matter atlas for consistent white matter tract parcellation across the lifespan.Neuroimage, 179:429–447

    2018

  49. [57]

    and Yang, Q

    Zhang, Y. and Yang, Q. (2018). An overview of multi-task learning.Natl. Sci. Rev., 5(1):30–43

    2018

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.