List of Papers Browse by Subject Areas Author List
Abstract
Brain connectivity emerges from multiple scales, from mesoscale axonal projections to macroscale white matter (WM) tract pathways. Virus tracing provides gold standard axonal projection at μm level but is restricted by limited virus injection locations and lack of streamlines. Diffusion MRI (dMRI) can cover brain-wide connectivity streamlines but cannot measure cellular level connections. Across-scales connectivity mapping including gold standard virus tracer injection, ex vivo ultra-high-resolution dMRI, and in vivo regular dMRI tractography offers a unique, complementary and possibly the only solution to elucidate the brain-wide mesoscale connectome, at the level of μm, 100 μm and 1000 μm respectively. However, this is hindered by the lack of advanced techniques to generate large-scale and high-throughput streamlines from virus tracing data and to bridge across-scales connectivity. To directly integrate connectivity across scales, we propose ACMap, a novel framework that reconstructs high-fidelity streamlines from virus tracing. ACMap leverages fast marching method (FMM) with a simple and interpretable speed function for continuous tracing of fiber pathways by casting streamline propagation as a front evolution process, where the interface advances progressively outward from the injection site. The proposed ACMap reconstructs high-fidelity and dense streamlines, enabling across-scales integration and validation of fiber pathways derived from dMRI and virus tracer data. Qualitative and quantitative assessments demonstrate that ACMap outperforms other competitive methods by a large margin. Further comparison between ACMap streamlines and dMRI tractography streamlines shows consistency of traced fibers. ACMap represents a significant step towards mapping across-scales connectivity of the brain, providing a new approach for accurate mapping of the brain-wide mesoscale connectome.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026/paper/4031_paper.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to the Code Repository
N/A
Link to the Dataset(s)
N/A
BibTex
@InProceedings{LinRun_ACMap_MICCAI2026,
author = { Lin, Runjia AND Zhu, Tianjia AND Ouyang, Minhui AND Huang, Hao},
title = { { ACMap: Across-scales Connectivity Mapping of Primate Brain } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 16894},
month = {September},
page = {pending}
}
Reviews
Review #1
- Please describe the contribution of the paper
The paper introduces ACMap, a novel computational framework for reconstructing dense and continuous axonal streamlines from virus tracing data. The proposed method formulates streamline reconstruction as a source-constrained shortest-path problem, solved using the fast marching method with a speed function derived directly from the tracer intensity. This allows the generation of high-resolution fiber trajectories from mesoscale tracer data, which can then be compared with dMRI tractography or used for integration of brain connectivity across multiple spatial scales. The work aims to bridge the gap between microscopy-based brain connectivity and macroscale imaging-derived connectivity, providing a potential framework for cross-scale validation of brain connectomics.
- Please list the major strengths of the paper: you should highlight a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.
- The paper tackles the challenging problem of reconstructing brain-wide structural connectivity from virus tracing data and linking it to dMRI-based tractography. Bridging connectivity across spatial scales is a key open problem in computational neuroscience.
- The authors formulate streamline reconstruction from tracer intensity as a continuous front propagation problem governed by the Eikonal equation and solved using FMM. This is an interpretable alternative to traditional neuron tracing methods that often rely on discrete skeletonization or heuristic graph-based approaches. More importantly, the proposed methods are claimed to operate fast at the scale of the whole brain, which is an important strength.
- Compared to existing neuron tracing methods (e. g. , NeuTube, APP2, Rivulet2), the method appears to better capture dense fiber bundles, which is important for mesoscale connectivity analysis.
- Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.
- A central concern is that there is no true ground-truth for the reported streamline reconstruction, yet the evaluation relies heavily on comparing reconstructed streamlines to the raw tracer intensity. Virus tracing data are known to be noisy and contain artifacts (e. g. , vasculature signal, non-specific labeling), which may not correspond to true axonal pathways. This directly affects: ○ The speed function assumption (intensity → axon likelihood), which may not hold in practice. ○ The Dice, which are computed against raw intensity and therefore may not reflect biologically meaningful accuracy.
- Additionally, the Dice score is computed on binarized maps, but the paper does not specify how the threshold is selected or how sensitive the results are to this choice. Since thresholding can significantly affect Dice scores of different methods, this raises concerns about reproducibility and the fairness of comparison.
- The method assumes that higher tracer intensity corresponds to higher axon likelihood and thus faster propagation. However, given noise, artifacts, and variability in tracer uptake, this assumption may not be robust and could lead to biased or spurious pathways.
- The paper lacks validation of whether reconstructed streamlines correspond to known anatomical targets. For example, the authors could validate whether distal projection regions match established connectivity patterns reported in prior tracing or connectivity studies. In the current format, it is unclear whether the method reconstructs biologically meaningful pathways or simply follows noisy intensity patterns.
- There is no clear visualization of: ○ Whole-brain reconstructed fibers ○ Regional streamlines overlaid on raw tracer intensity ○ Direct comparison overlays between ACMap, baseline methods, and dMRI Such visualizations are critical for assessing the fidelity and anatomical plausibility of the results.
- Please rate the clarity and organization of this paper
Good
- Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.
The submission does not mention open access to source code or data, but provides a clear and detailed description of the algorithm to ensure reproducibility.
- Based on your review and your understanding of the MICCAI Scientific Code of Ethics, do you believe this submission may involve a potential ethics concern or violation?
N/A
- Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2026/en/REVIEWER-GUIDELINES.html
Minor comments
- In Fig. 6a, some dMRI tractography streamlines appear to extend outside the brain boundary, which raises concerns about preprocessing, masking, or tractography constraints. This should be clarified and explained, and what method was used.
- The method critically depends on the injection site as the source of front propagation, but it is unclear how this is determined: ○ Is it manually annotated? ○ Derived from atlas registration and a known region? ○ Based on intensity thresholding? Variability in injection site localization (e. g. , due to stereotaxic differences) could significantly affect the results, and this is not discussed.
- Figures are small and lack sufficient detail, making it difficult to assess the results and compare methods effectively.
- No quantitative runtime analysis or computational comparison with baseline methods is provided, although the authors repeatedly emphasize that the proposed method is fast. Without reporting execution time, scalability, or computational complexity in practice, it is difficult to assess whether the method offers a meaningful advantage in efficiency, especially for large volumetric datasets.
- Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making.
(2) Reject — should be rejected, independent of rebuttal
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
The evaluation of the method and comparison to baselines (which lack ground truth) is limited. Some key methodological details are missing for the evaluation (eg. injection site generation), and important visualizations are lacking.
- Reviewer confidence
Confident but not absolutely certain (3)
- [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.
N/A
- [Post rebuttal] Please justify your final decision from above.
N/A
Review #2
- Please describe the contribution of the paper
Providing a sound mapping of the brain’s connectivity is important to understand brain functioning. This work aims at elucidating the correspondence between results of tracer-based connectivity mapping at microscopic resolution with dMRI-based tracking at the mm level. Streamlines are reconstructed from 3D high-resolution images by (1) solving the eikonal equation using a fast-marching method, (2) computing a time-of-arrival map, and (3) tracking the streamline using gradient descent. Results are compared with other approaches for neuron tracing.
- Please list the major strengths of the paper: you should highlight a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.
The topic of this submission is within the scope of this conference. The text is straightforward to understand for a reader with knowledge in recent approaches for image analysis. The work described here is interesting and provides a demonstrated and considerable advance over previous approaches.
- Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.
A minor issue: Apparently, the letter \mu is missing in several locations, which led to some confusion. It is also suggested to use either italic or regular letters for units consistently.
- Please rate the clarity and organization of this paper
Good
- Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.
The submission does not mention open access to source code or data, but provides a clear and detailed description of the algorithm to ensure reproducibility.
- Based on your review and your understanding of the MICCAI Scientific Code of Ethics, do you believe this submission may involve a potential ethics concern or violation?
N/A
- Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2026/en/REVIEWER-GUIDELINES.html
N/A
- Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making.
(5) Accept — should be accepted, independent of rebuttal
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
Interesting method, demonstrated advance
- Reviewer confidence
Very confident (4)
- [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.
N/A
- [Post rebuttal] Please justify your final decision from above.
N/A
Review #3
- Please describe the contribution of the paper
This work proposes a novel fiber reconstruction method named ACMap for virus tracing data. Comparative experiments with competing methods, as well as validation against dMRI tractography results, demonstrate the favorable performance of the proposed approach.
- Please list the major strengths of the paper: you should highlight a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.
The ACMap method achieves excellent tractography reconstruction performance.
- Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.
Evaluations on more diverse datasets are still lacking. Additional tests across multiple independent datasets would further verify the strong robustness of this method and enhance its credibility.
- Please rate the clarity and organization of this paper
Satisfactory
- Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.
The submission does not mention open access to source code or data, but provides a clear and detailed description of the algorithm to ensure reproducibility.
- Based on your review and your understanding of the MICCAI Scientific Code of Ethics, do you believe this submission may involve a potential ethics concern or violation?
N/A
- Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2026/en/REVIEWER-GUIDELINES.html
N/A
- Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making.
(5) Accept — should be accepted, independent of rebuttal
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
Comparative experiments with competing methods, as well as validation against dMRI tractography results, demonstrate the favorable performance of the proposed approach (ACMap). The overall writing is well-organized without significant deficiencies.
- Reviewer confidence
Very confident (4)
- [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.
Accept
- [Post rebuttal] Please justify your final decision from above.
Answered my concerns.
Author Feedback
We thank the reviewers for their constructive comments. We summarized the main critiques into 5 categories and provided responses with Rx-y referring to the response to comment #y from reviewer #x.
Category 1: Lack of biological ground-truth and unvalidated assumption R1-1 and R1-3: We respectfully note that, despite the fact that no method is an “absolute” gold standard for tracing axonal pathways, among all the tracing techniques, virus tracing is widely regarded as one of the closest experimental proxies for axonal connectivity (e.g. Refs 1,12). It has also been uniformly considered as gold standard for the diffusion-MRI-based tractography method evaluation (e.g. Refs. 1, 8,10, 12). Beyond MR field, the more general basic neuroscientific field endorses the validity of virus tracing, as demonstrated by Allen Brain Atlas mouse connectome (Wang et al., Cell 2020; PMID: 32386544). The virus-tracing dataset adopted in this study had already undergone rigorous artifact mitigation by the original providers (Ref. 17), with this artifact-removal procedure further enhanced by multiple steps in ACMap.
Category 2: Evaluation and robustness R1-1 and R1-2: We followed a widely used evaluation method of setting a voxel to 1 if traversed by at least one streamline and 0 otherwise (e.g. Wakana et al., Neuroimage 2007; PMID: 17481925) for Dice score evaluation, and applied this evaluation to ACMap and all competing methods. We also reported AUC (Fig. 5), a threshold-independent metric (Ref. 12). We will clarify the evaluation above in the revised manuscript.
Category 3: Validation and visualization R1-4: The reconstructed commissural and projection fibers align with the pathways well characterized and delineated in a histology-based marmoset atlas book (Paxinos et al., 2025). Specifically, commissural fibers matched canonical interhemispheric tracts; and projection fibers followed atlas-defined trajectories from cortex through the internal capsule and cerebral peduncle.
Paxinos, G., et al. (2025). The marmoset brain in stereotaxic coordinates. Academic press.
R1-5 and R1-8: For clearer visualization, we will generate a figure including enlarged streamlines overlaid on raw virus-tracer intensity maps for ACMap and all competing methods included in the paper in GitHub repository. Understanding the limitation of limited injection site of adopted marmoset virus tracing dataset, we will also extensively and rigorously validate ACMap using Allen Institute mouse virus-tracing data with thousands of injection sites (Ref. Wang et al., Cell 2020).
R1-7: The injection site was already delineated by the original dataset providers (Ref. 17) using a rigorous method to localize and detect labeled cell bodies with intensity above a certain threshold as the injection site. Despite robustness of injection site localization for virus tracing, we agree for validation of other fiber tracing method, precise registration of injection site (translated to fiber tractography initialization location) is critical and will be discussed in the revised manuscript.
Category 4: Computational efficiency R1-9: We respectfully clarify a possible misreading: the manuscript does not claim computational speed as a contribution. “Fast Marching Method” is the standard name of a well-established algorithm (Ref. 15), not a statement about ACMap’s runtime performance. We will add a runtime analysis and discuss that GPU acceleration could substantially improve ACMap’s efficiency in the revised manuscript.
Category 5: Minor concerns R1-6 and R1-8: In Fig. 6a, the dMRI streamlines appear to exit the brain due to a surface-rendering artifact. In the original 3D volume, the streamlines remain within the brain mask, and tractography was performed following Ref. 22.R2-1: We will restore the missing µ symbols and harmonize the units throughout the manuscript. R4-1: We agree that evaluation across multiple datasets would further strengthen the study. We will note this for future work.
Meta-Review
Meta-review #1
- Your recommendation
Invite for Rebuttal
- Please justify your decision. In case you deviate from the reviewers’ recommendations, explain in detail the reasons why. In case of an invitation for rebuttal, clarify which points are important to address in the rebuttal.
Two reviewers support acceptance and acknowledge the methodological novelty and the meaningful cross-scale value of ACMap’s brain connectivity research, whereas one reviewer raises substantial critical limitations. To resolve the conflicting reviews, the authors must fully address the key critical points raised by Reviewer 1, including the lack of biological ground truth, unreliable evaluation based on noisy virus tracing signals, unclear threshold setting for Dice metrics, unvalidated assumption between tracer intensity and axon likelihood, ambiguous injection site localization, insufficient anatomical validation and comprehensive visualizations, as well as missing computational efficiency analysis. Minor concerns, such as limited dataset diversity, formal notation inconsistencies and problematic dMRI preprocessing artifacts, should also be addressed concisely in the rebuttal.
- After you have reviewed the rebuttal and updated reviews, please provide your recommendation based on all reviews and the authors’ rebuttal.
Reject
- Please justify your recommendation.
After reviewing the initial feedback, the author’s rebuttal, and the post-rebuttal comments, I recommend rejecting this paper for several reasons. The main issue is the lack of independent biological ground truth for the streamline reconstruction task. The evaluation relies solely on comparing reconstructed streamlines with the raw tracer data used to generate them, creating a circular validation. Although virus tracing offers insights into brain connectivity, it has significant noise and artifacts that can lead to misleading reconstructions. The authors claim that higher tracer intensity indicates a higher likelihood of axons, but they provide no evidence to support this assumption. Furthermore, they assert that reconstructed fibers align with known pathways in the Paxinos marmoset brain atlas but provide no corroborating evidence, such as visualizations or quantitative comparisons. The promise of improved visualizations post-acceptance does not compensate for the lack of validation in the manuscript. The authors also fail to clarify how the threshold for Dice score calculations was chosen, affecting the reliability of their comparative evaluations. Additionally, key methodological details, such as the injection site localization and computational efficiency analysis, remain unclear. The rebuttal focuses on commitments for future improvements rather than evidence from the submitted manuscript. Acceptance should be based solely on the current submission, not on promises of additional work.
Meta-review #2
- After you have reviewed the rebuttal and updated reviews, please provide your recommendation based on all reviews and the authors’ rebuttal.
Accept
- Please justify your recommendation.
I think the authors did a good job addressing the concerns of Rev. 1 in the rebuttal. Unfortunately, ground truths are not available for tractography. The authors proposed a robust method for predicting streamlines from virus-tracing data, which is a step forward in developing silver standards for assessing tractography methods.
Meta-review #3
- After you have reviewed the rebuttal and updated reviews, please provide your recommendation based on all reviews and the authors’ rebuttal.
Accept
- Please justify your recommendation.
All reviewers recommended acceptance. The authors have adequately addressed the reviewers’ questions and concerns. The paper is therefore recommended for acceptance.