List of Papers Browse by Subject Areas Author List
Abstract
We present a real-time visualization system for Transcranial Magnetic Stimulation (TMS), a non-invasive neuromodulation technique for treating various brain disorders and mental health diseases. Our solution targets the current challenges of slow and labor-intensive practices in treatment planning. Integrating Deep Learning (DL), our system rapidly predicts electric field (E-field) distributions in 0.2 seconds for precise and effective brain stimulation. The core advancement lies in our tool’s real-time neuronavigation visualization capabilities, which support clinicians in making more informed decisions quickly and effectively. We assess our system’s performance through three studies: First, a real-world use case scenario in a clinical setting, providing concrete feedback on applicability and usability in a practical environment. Second, a comparative analysis with another TMS tool focusing on computational efficiency across various hardware platforms. Lastly, we conducted an expert user study to measure usability and influence in optimizing TMS treatment planning. The system is openly available for community use and further development on GitHub: https://github.com/lorifranke/SlicerTMS.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2024/paper/2402_paper.pdf
SharedIt Link: https://rdcu.be/dV5yu
SpringerLink (DOI): https://doi.org/10.1007/978-3-031-72089-5_54
Supplementary Material: https://papers.miccai.org/miccai-2024/supp/2402_supp.zip
Link to the Code Repository
https://github.com/lorifranke/SlicerTMS
Link to the Dataset(s)
N/A
BibTex
@InProceedings{Fra_SlicerTMS_MICCAI2024,
author = { Franke, Loraine and Luo, Jie and Park, Tae Young and Kim, Nam Wook and Rathi, Yogesh and Pieper, Steve and Ning, Lipeng and Haehn, Daniel},
title = { { SlicerTMS: Real-Time Visualization of Transcranial Magnetic Stimulation for Mental Health Treatment } },
booktitle = {proceedings of Medical Image Computing and Computer Assisted Intervention -- MICCAI 2024},
year = {2024},
publisher = {Springer Nature Switzerland},
volume = {LNCS 15006},
month = {October},
page = {575 -- 585}
}
Reviews
Review #1
- Please describe the contribution of the paper
The paper provides an interesting software integration of a recently developed deep learning-based E-field prediction algorithm with a widely used open-source software (3D Slicer) for real-time visualization. The primary contribution is to enhance the practicality of TMS by providing a real-time E-field visualization toolbox, which is of clinical significance. This toolbox, SlicerTMS, features 1) fast prediction and visualization speed, 2) capability of predicting from real-time coil placement, and 3) possibility of coil interaction through AR and VR devices.
- 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.
Real-time (or near real-time, depending on the local or remote GPU specs) E-field visualization has great usability in practice. In this paper, it is achieved by applying the recent development in DL models (such as 3D-ResUnet) for fast E-field prediction. It also seems to be possible to upgrade the prediction algorithm in the future. Combined with the efficient use of 3D Slicer software, this new toolbox achieves significant speed-up compared to non-DL based visualization tools such as SimNIBS.
- 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 paper does not sufficiently report on the network structure and the evaluation metrics for model accuracy, which limits the reproducibility and significance of the results. It is a common trade-off between the speed and the accuracy of the E-field prediction algorithm. Therefore, despite the speed advantage, it would be helpful to provide reproducible, complete reports on model accuracy, so that readers can better decide the practical usability of the proposed toolbox.
There is a lack of methodology novelty compared the the prior works such as [31], where the prediction algorithm was proposed. A direct comparison or clarification would help in establishing the paper’s contribution more distinctly.
The precision of the predictions and the overall effectiveness of the supervised learning approach depend on the quality of training data, the specifics of the network structure, as well as coil type used in training and prediction and the precision of pre-trained simulation data. Although it is possible to improve with an upgrade of the back-end prediction algorithm, these factors are not adequately addressed in the paper, which could be informative for the readers and practitioners.
- Please rate the clarity and organization of this paper
Very 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 has provided an anonymized link to the source code, dataset, or any other dependencies.
- Do you have any additional comments regarding the paper’s reproducibility?
The paper provides an anonymized link to the source code of the software and the examples. However, the prediction model used by the software is difficult to be reproduced, because it does not clarify network structure details or if the prediction model is identical to a prior work.
- Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review. Pay specific attention to the different assessment criteria for the different paper categories (MIC, CAI, Clinical Translation of Methodology, Health Equity): https://conferences.miccai.org/2024/en/REVIEWER-GUIDELINES.html
The reported numerical evidence (e.g. both of the two reported tables) focuses on the computational speed. This is already convincing based on the report of prior works (0.2 second for prediction and visualization in this paper vs 0.24 second for prediction in the prior work). More supportive evidence on maintaining similar precision despite the additional speed-up is welcomed.
- 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
Weak Accept — could be accepted, dependent on rebuttal (4)
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
This work clearly demonstrates potential clinical significance and establishes its usability with multiple real-world applications. However, it is lacking in the report of model accuracy, which is not robustly validated against appropriate performance metrics. The paper would be stronger if it is justified that the cost in model accuracy for speed is fully negligible or acceptable.
- Reviewer confidence
Very confident (4)
- [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed
N/A
- [Post rebuttal] Please justify your decision
N/A
Review #2
- Please describe the contribution of the paper
The authors propose a two-part system for TMS neuro-navigation. The first part is a tradition SlicerIGT-based system with optimical tracking of the TMS stimulator and the patient. The second part is a deep-learning based simulation of the electric field. This is based off of learning a particular EM simulation, using the network to perform this approximately and much more quickly. The two together lead to a visualisation interface which can be used during TMS interventions. The system is validated with 10 subjects and 4 expert users. The interfaces include a standard Slicer instance and also an AR interface.
- 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.
- There is a lot of good material here - the motivation is clear, the system well-designed, the videos convincing. Unfortunately, MICCAI page limits have meant some really good material is relegated to the Supplementary material (i.e. the user study results)
- Solid statistical testing despite the relatively few number of patients in the study.
- 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.
- Missing information about patient tracking. One assumes that you calibrate the relationship between the patient and their sensor using facial landmarks, but this should be made explicit.
- Could have performed an ablation user study (i.e. only the navigation components with no simulation ones) as to have a comparative method for the user study. This would help to decouple the utility of the two distinct components.
- Please rate the clarity and organization of this paper
Excellent
- 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.
- Do you have any additional comments regarding the paper’s reproducibility?
The framework is conceptually clear, although a bit more could have been done to render the network architecture components more clearly in the paper or supplementary material.
- Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review. Pay specific attention to the different assessment criteria for the different paper categories (MIC, CAI, Clinical Translation of Methodology, Health Equity): https://conferences.miccai.org/2024/en/REVIEWER-GUIDELINES.html
- Table 1 - bolding is a bit inconsistent given some rows don’t have any bolded results (Subj 3 vis.). Ultimately, given that this is a hardware comparison rather than algorithmic comparison, I don’t think the bolding is particularly necessary.
- Figure 4 should be augmented with a visualisation of the SlicerTMS simulation but over the cortex and without the TMS stimulator obscuring the view (as with the SimNIBS visualisation).
- Suppl. Material: Figure 4 part 2 really doesn’t need to be a box-and-whisker plot, only having four points in each set. I would just show the points alone.
- Some minor typesetting issues such as ’figure 8’ on page 3 (note incorrect opening single quote mark)
- 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
Accept — should be accepted, independent of rebuttal (5)
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
This is a good example of a real MIC & CAI paper. The CAI components are state-of-the-art for an under-investigated intervention. The MIC components may not necessarily be the most theoretically novel but they are well-placed and well-motivated here.
- Reviewer confidence
Very confident (4)
- [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed
N/A
- [Post rebuttal] Please justify your decision
N/A
Review #3
- Please describe the contribution of the paper
They present SlicerTMS, an open-source software enabling real-time E-field prediction and visualization for TMS treatment on the Slicer’s platform, using deep learning (DL) within a neuronavigation system for immediate visualization .
- 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.
It is an innovative project, the paper is well written. It enhance the Slicer platform with a new clinically useful tool.
- 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.
Would have been interesting to have more than 1 subject scanned and tried using the prototype to make sure it can be generalized.
- Please rate the clarity and organization of this paper
Very 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 has provided an anonymized link to the source code, dataset, or any other dependencies.
- Do you have any additional comments regarding the paper’s reproducibility?
Using Slicer which is an open source platform, and SlicerTMS seems to be available on a GitHub (for 7 mo).
- Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review. Pay specific attention to the different assessment criteria for the different paper categories (MIC, CAI, Clinical Translation of Methodology, Health Equity): https://conferences.miccai.org/2024/en/REVIEWER-GUIDELINES.html
Really interesting paper, well written, with a direct clinical application.
- 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
Accept — should be accepted, independent of rebuttal (5)
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
Well written, interesting project with clinical impact.
- Reviewer confidence
Confident but not absolutely certain (3)
- [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed
N/A
- [Post rebuttal] Please justify your decision
N/A
Author Feedback
N/A
Meta-Review
Meta-review not available, early accepted paper.