Fostering High School Careers in Brain Mapping Research – Exvivo MRI Brain Vessel Segmentation
Poster No:
588
Submission Type:
Abstract Submission
Authors:
Erendira Garcia Pallares1, Rogeny Herisse1, Cole Analoro1, Caroline Magnain1, Chiara Mauri1, Julie Price1, Bruce Fischl1, Patrick Hof2, Carmen Feire Cobo2, Andre Vanderkouwe1, Divya Varadarajan1, Robert Frost1, Nam Tran1, Stephanie Lin1, Itzel Garcia1, Priyanka Onta1, Aliyah Jama1
Institutions:
1Massachusetts General Hospital, Charlestown, MA, 2Mount Sinai Hospital, New York, NY
First Author:
Co-Author(s):
Introduction:
Deep learning-driven imaging analysis has achieved unparalleled accuracy in tasks like classification, segmentation, and image synthesis (Li, Jiang, Zhang, & Zhu, 2023). Manual segmentation remains essential to image analysis because it provides the ground truth labeled images needed to develop, optimize and evaluate semi-automatic and fully automatic segmentation techniques (Haque & Neubert, 2020). In this study, we highlight the contributions of high school interns who assisted with manual segmentation to improve vessel identification in ex vivo MRI brain images. Their work plays a crucial role in creating a detailed whole-brain vascular map and inter-individual variability in ex vivo MRI images. Additionally, the project offered an educational experience for the interns, including talks from neuroanatomists and clinicians on brain vessel structure and function, as well as exposure to programming and machine learning in neuroscience.
Methods:
Internship Pedagogy
Interns participate in a 10-week program, Monday to Friday from 4 PM to 6 PM, with an optional independent work period from 6 PM to 7 PM. The first two weeks (Figure 1) focus on intensive training (10-15 hours/week), including: 1) safe research practices, 2) lectures on labeling techniques and protocols, 3) workshops on neuroanatomy, deep learning, programming, and MRI imaging, and 4) professional development through seminars and one-on-one mentorship. All sessions are led by faculty and research staff from Massachusetts General Hospital, Boston University, and Mount Sinai. At the end of the program, interns present their projects and key takeaways in a formal poster session to MGH staff.
Vessel Segmentation
Each intern was assigned two 64³ mm patches from randomized regions of the brain's cerebral hemisphere (Figure 2) from ex vivo MRI data (7T, 120 µm) for labeling. For each patch, both dark and bright vessels were labeled using Freeview, a visualization tool within the FreeSurfer package. Interns utilized all three orthogonal planes to cross-reference voxel locations between planes. They were also provided with various histological brain schematics to guide their labeling process. Additionally, interns received weekly check-ins with the internship coordinator, who provided feedback and suggested edits for each labeled case.
Interns participate in a 10-week program, Monday to Friday from 4 PM to 6 PM, with an optional independent work period from 6 PM to 7 PM. The first two weeks (Figure 1) focus on intensive training (10-15 hours/week), including: 1) safe research practices, 2) lectures on labeling techniques and protocols, 3) workshops on neuroanatomy, deep learning, programming, and MRI imaging, and 4) professional development through seminars and one-on-one mentorship. All sessions are led by faculty and research staff from Massachusetts General Hospital, Boston University, and Mount Sinai. At the end of the program, interns present their projects and key takeaways in a formal poster session to MGH staff.
Vessel Segmentation
Each intern was assigned two 64³ mm patches from randomized regions of the brain's cerebral hemisphere (Figure 2) from ex vivo MRI data (7T, 120 µm) for labeling. For each patch, both dark and bright vessels were labeled using Freeview, a visualization tool within the FreeSurfer package. Interns utilized all three orthogonal planes to cross-reference voxel locations between planes. They were also provided with various histological brain schematics to guide their labeling process. Additionally, interns received weekly check-ins with the internship coordinator, who provided feedback and suggested edits for each labeled case.
·Training Structure for Weeks 1–2 of the Laboratory of Computational Neuroimaging High School Internship
·Example patches selected from different regions of two ex vivo MRI hemispheres (120 µm)
Results:
For each patch, two interns independently labeled the vessels, and their combined labeling was used as the final ground truth to help minimize missed vessels. The patches were then divided into a validation set of 3 patches and a test set of 4 patches. The method was applied to the test patches, and the results were compared to the ground truth using the Dice score and Hausdorff distance (Huttenlocher, Klanderman, & Rucklidge, 1993). The proposed method achieved a Dice score of 0.491.
Conclusions:
Manual segmentation remains essential for providing ground truth data that refines these methods. This study highlights the contributions of high school interns in manual segmentation, which were crucial for creating an accurate whole-brain vascular map and advancing our understanding of brain anatomy and vascularity. The method can also be extended to other regions of interest, such as subcortical structures like the hippocampus and thalamus, contributing to the development of a comprehensive brain atlas.
Education, History and Social Aspects of Brain Imaging:
Education, History and Social Aspects of Brain Imaging 1
Modeling and Analysis Methods:
Image Registration and Computational Anatomy
Segmentation and Parcellation 2
Neuroinformatics and Data Sharing:
Brain Atlases
Novel Imaging Acquisition Methods:
Anatomical MRI
Keywords:
Acquisition
Atlasing
Computational Neuroscience
Data Registration
MRI
STRUCTURAL MRI
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
Rizwan I Haque, I., & Neubert, J. (2020). Deep Learning Approaches to Biomedical Image segmentation. Informatics in Medicine Unlocked, 18, 100297. https://doi.org/10.1016/j.imu.2020.100297
Huttenlocher, D. P., Klanderman, G. A., & Rucklidge, W. J. (1993). Comparing images using the Hausdorff distance. IEEE Transactions on Pattern Analysis and Machine Intelligence, 15(9), 850–863. https://doi.org/10.1109/34.244672