Paper Info Reviews Meta-Review Author Feedback Post-rebuttal Meta-Reviews

Authors

Ye Wu, Yoonmi Hong, Sahar Ahmad, Pew-Thian Yap

Abstract

Most existing diffusion tractography algorithms are affected by gyral bias, causing the termination of streamlines at gyral crowns instead of sulcal banks. In this paper, we propose a tractography technique, called active cortex tractography (ACT), to overcome gyral bias by enabling fiber streamlines to curve naturally into the cortex. We show that the cortex can play an active role in cortical tractography by providing anatomical information to overcome orientation ambiguities as the streamlines enter the superficial white matter in gyral blades and approach the cortex. This is achieved by devising a direction scouting mechanism that takes into account the white matter surface normal vectors. The scouting mechanism allows probing of directions further in space to prepare the streamlines to turn at appropriate angles. The surface normal vectors guide the streamlines to turn into the cortex, perpendicular to the white-gray matter interface. Evaluation using synthetic, macaque and human data with different streamline seeding schemes demonstrates that ACT improves cortical tractography.

Link to paper

DOI: https://doi.org/10.1007/978-3-030-87234-2_44

SharedIt: https://rdcu.be/cyl8G

Link to the code repository

N/A

Link to the dataset(s)

N/A

Reviews

Review #1

Please describe the contribution of the paper

In this paper, the authors proposed a novel approach to overcome the known gyral bias of tracking algorithms in tractography, leveraging two main concepts: asymmetric fiber orientation distributions (FODs), to further tailor the tracking procedure at the subvoxel level; and surface-based priors, taking advantage of the known boundaries between grey and white matter. Based on both synthetic and real datasets, the results show streamlines with wider angle profiles when approaching the cortex.
Please list the main strengths of the paper; you should write about 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 authors leverage two existing and relatively novel approaches, asymmetric FODs and surface-based priors (citing in both cases relevant papers, refs. 11 and 13), to tackle the issue of gyral bias in tractography; the surface-based priors are something that it is not explored as much as it should - this paper makes a perfect case for it;
- The overall approach is described in a synthetic but clear and direct way;
- The authors put a remarkable effort in validating in both synthetic and real data, using a comparison both a probabilistic algorithm and a deterministic one;
Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.
- The results are mostly qualitative, but this is appropriate as quantitative validation in tractography is nontrivial;
- Some notes on the implementation could have been interesting, but understandably there is already a lot on their plate;
Please rate the clarity and organization of this paper

Very Good
Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance

The authors relied on a synthetic dataset (not provided but generated through an open-source tool and with relevant details) and on real ones (HCP datasets, publicly available). In terms of methods, they are described in mathematical terms (and with clean and precise formalism) but in terms of implementation few details are provided.
Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://miccai2021.org/en/REVIEWER-GUIDELINES.html
- I would suggest to pick a different acronym or shorter name for the approach, as ACT is widely used to refer to the anatomically-constrained tractography approach developed in MRtrix (https://mrtrix.readthedocs.io/en/latest/quantitative_structural_connectivity/act.html);
- The mathematical formalism is very clear, but I think there will be some benefit in adding a couple of sentences further explaining the asymmetric FODs and surface prior approaches - I’m mentioning this because I think it could be done without much issue just resizing a little bit the figures;
- It would be better to specify which deterministic algorithm has been used (the one shortened DT), as there are more than one deterministic algorithm, each with different advantages and disadvantages;
- Another interesting aspect that would deserve a couple of sentence in the results (if some space is to be saved by figure resizing) is the potential implication in white matter for this approach: figure 7 shows how the CST looks like, but I would be very curious to see how two intersecting bundle are affected by the asymmetric FODs; of course this is outside the scope of this work, but a comment on potential advantages there would be interesting;
Please state your overall opinion of the paper

accept (8)
Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

The paper is interesting and tackles a current issue of tractography. The methods are well described and the provided results are more than appropriate to showcase the approach.
What is the ranking of this paper in your review stack?

1
Number of papers in your stack

5
Reviewer confidence

Very confident

Review #2

Please describe the contribution of the paper

The paper proposed a novel tractography algorithm using anatomical priors to reduce the gyral bias. It uses the normal to the white matter surface to guide the streamline propagation near the cortex. The method also uses asymmetric fibre ODF to better capture the fibre orientations. The experimentation is done on synthetic, animal and human dataset. Results show the method provides better cortical coverage than conventional tractography methods.
Please list the main strengths of the paper; you should write about 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 manuscript is clearly written and interesting.

The proposed method addresses one of the challenges of tractography methods.

The manuscript proposes multiple experiments showing the benefits of the proposed method over conventional tractography methods. In particular using synthetic, monkey and HCP data.
Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.

The manuscript acknowledges the existence of other method using anatomical priors, but don’t discuss them regarding cortical tractography. In particular, authors stated “Unlike existing tractography methods, ACT is designed with cortical tractography in mind”. The method SET (St-Onge et al., 2018), referenced in the manuscript, is designed to reduce the giral bias, also using normals to the surface to guide the tractography. Similarly, the method proposed by Teillac et al. (2017; DOI:10.1007/978-3-319-66182-7_61) uses the surface to guide the tractography in region close the the cortex. The manuscript would benefit from discussing how the proposed method uses the surface information in contrast with previously proposed methods.

The authors studied the effect of spatial resolution of the cortical coverage. However, the experimentation is limited to synthetic data. The conclusion would have been stronger if this was performed on the monkey data as well (by artificially lowering the resolution). For instance, by reporting the results of Fig.5. at various resolutions.
Please rate the clarity and organization of this paper

Very Good
Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance

The code of the proposed method is not available. The synthetic data used to test the method is not available. However, similar synthetic data could be generated using the Phantomas toolbox. Authors used publically available monkey and human data for their experiments.
Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://miccai2021.org/en/REVIEWER-GUIDELINES.html

p2. There is no reference for the AFODF method. Please describe how the AFODF were obtained. Was it done identically for the synthetic, monkey and human data?

p3. Fig.1 (left) is unclear to me. Please describe the elements of the figure.

p4. Was the DT from the MRtrix3 software package? Please add the reference for the implementation used.

p4. How many seed/streamlines were used (in particular for the synthetic experiment)?

p6. Fig.4. Although the figure provides an interesting visualization contrast between the tested methods, it misses the opportunity to compare against the ground truth curvature map available for this synthetic dataset. Could the authors include a subfigure showing the ground truth curvature of the trajectories of the fanning phantom? Moreover, the paper would benefit from quantifying the differences to the ground truth for the tested methods.

p.6. How was defined the stopping criterion of the tractography algorithms (e.g. the tracking mask)? The authors stated they tested their method using GM seeding. How was it done? Wouldn’t the streamlines all stop at the seed location?

p.6. Please add the identification code of HCP dataset used.

p.6. Please specify how surfaces were obtained for the monkey and human data.

p8. Table 1 is not reference in the text. Please describe. How were the gyral crown and sucal wall parcellated?
Please state your overall opinion of the paper

Probably accept (7)
Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

The paper proposes an interesting method to improve tractography using anatomical priors. The experimentation is sound, but misses quantitative analysis against ground truth data. The method is adequately compared to conventional tractography algorithm, but little discussion is offered on state-of-the-art method designed to reduce the gyral bias.
What is the ranking of this paper in your review stack?

2
Number of papers in your stack

3
Reviewer confidence

Very confident

Review #3

Please describe the contribution of the paper

This submission describes a tractography method called active cortex tractography. Synthetic, macaque and human diffusion weighted imaging data were used to demonstrate the cortical tractography using a directional scouting mechanism. The method tries to assess the gyral bias that can be present in superficial white matter.
Please list the main strengths of the paper; you should write about 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.
- Novel method to perform a more reliable tracking of superficial white matter.
- Fanning and bending configurations are challenging in diffusion tensor estimation and can lead to false interpretations. Therefor this novel active cortex tractography method has added value.
- HCP data were used in the analysis. This is a well-known dataset and allows for replication.
- The newly proposed method seems to be robust in showing the average curvature of streamlines.
Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.
- In the field of diffusion weighted imaging the abbreviation ACT is used to describe Anatomically-Constrained Tractography. This is widely used and originates from the software package MRTrix. Hence the abbreviation should be changed. It can be questioned to what extend this method is “active”, as it has nothing to do with functional activation.
- It is difficult to tell which method is best and there is no ground truth to compare against. It indeed seems that the active cortex tractography might perform better (compared to iFOD2 and DT), but what about other algorithms?
Please rate the clarity and organization of this paper

Good
Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance

Code and data were not provided.
Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://miccai2021.org/en/REVIEWER-GUIDELINES.html

The comparison between different methods is interesting. Though it can be questioned to what extend it is relevant to use a voxel size of six or four millimeter, since the spatial resolution of DWI is usually in the order of two millimeters or smaller. The simulated data shows nicely the curvature difference.
Please state your overall opinion of the paper

borderline accept (6)
Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

It is difficult to tell which method is best and there is no ground truth to compare against. It indeed seems that the active cortex tractography might perform better (compared to iFOD2 and DT), but what about other algorithms?
What is the ranking of this paper in your review stack?

2
Number of papers in your stack

3
Reviewer confidence

Somewhat confident

Primary Meta-Review

Please provide your assessment of this work, taking into account all reviews. Summarize the key strengths and weaknesses of the paper and justify your recommendation. 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.

Reviewers agree that this is a well-written paper on a sufficiently novel method for a relevant problem. Authors are advised to pick a different name (or, at least, acronym), since ACT is already taken for “anatomically constrained tractography”. A weakness is that there is no comparison to previous methods that have used the cortical surface to guide tractography. Authors should at least argue more clearly how their approach differs from those existing ones. Individual reviews contain several other minor comments that should be addressed in the final version. Despite this, there was no doubt that this paper should be presented at MICCAI 2021.
What is the ranking of this paper in your stack? Use a number between 1 (best paper in your stack) and n (worst paper in your stack of n papers).

2

Author Feedback

Comparison with previous methods guided by cortical surface: Instead of fiber tracking using dMRI near the white-matter surface, SET (St-Onge et al., 2018) uses a surface flow technique to simulate artificial streamlines that intersect orthogonally with the white-matter surface. However, it is unclear whether the fiber streamlines enter the sulcal walls orthogonally. Instead of relying on artificial ‘surface flow’, our method considers both the dMRI signal near the cortical surface and the surface geometry, Our method does not rely on artificial simulation from higher-resolution structural images (Teillac et al., 2017). Moreover, our method has demonstrated outstanding performance even with deterministic tractography, hence avoiding the computational burden associated with global tractography (Teillac et al., 2017).

“Active” and functional activation: In this paper, the term ‘active’ is referring to the fact that the cortical surface is playing an active role in tractography and has nothing to do with functional activation in fMRI.

back to top

Active Cortex Tractography