aiqbench analysis
This study provides a rigorous systematic evaluation of Transformer and RNN architectures for diffusion MRI tractography, addressing the critical challenge of balancing local diffusion data with global anatomical plausibility. By introducing a novel generation-validation phase, the authors demonstrate superior performance on benchmark datasets while quantifying model sensitivity to noisy training streamlines and missing bundles. These findings offer essential insights into the practical limits and optimization strategies for sequence-based deep learning in neuroimaging reconstruction tasks.
A foundation for systematic analysis of transformers and RNNs for tractography
arXiv:2610.01894v1 Announce Type: new Abstract: Machine learning (ML) has emerged as a promising approach for improving diffusion MRI (dMRI) tractography, a task that remains limited by the intrinsic tension between local diffusion information and global anatomical plausibility. In this work, we systematically evaluate recurrent neural networks (RNNs) and Transformer models for iterative tractography, with particular attention to training strategies, input representations (including convolutional neural network (CNN)-based embeddings and end-of-sequence (EOS) tokens), and hyperparameter selection. We introduce a generation-validation phase enabling supervision at the streamline level during training, allowing supervision despite the mismatch between local loss functions and global streamline quality. Using the ISMRM2015 tractography challenge dataset, our models achieve the highest reported performance to date. Through controlled experiments, we quantify the impact of missing bundles, noisy or imperfect training streamlines, and invalid fibers in the training set. Finally, we demonstrate the applicability of our best-performing models for in vivo data from the Tractoinferno database. Overall, our results highlight both the potential and the limits of sequence-based deep learning models such as Transformers and RNNs for tractography, and emphasize the need for improved phantoms and evaluation methods for in vivo validation. We provide takeaways and recommendations for future researchers training and validating sequence-based supervised methods for tractography.
Abstract:Machine learning (ML) has emerged as a promising approach for improving diffusion MRI (dMRI) tractography, a task that remains limited by the intrinsic tension between local diffusion information and global anatomical plausibility. In this work, we systematically evaluate recurrent neural networks (RNNs) and Transformer models for iterative tractography, with particular attention to training strategies, input representations (including convolutional neural network (CNN)-based embeddings and end-of-sequence (EOS) tokens), and hyperparameter selection. We introduce a generation-validation phase enabling supervision at the streamline level during training, allowing supervision despite the mismatch between local loss functions and global streamline quality. Using the ISMRM2015 tractography challenge dataset, our models achieve the highest reported performance to date. Through controlled experiments, we quantify the impact of missing bundles, noisy or imperfect training streamlines, and invalid fibers in the training set. Finally, we demonstrate the applicability of our best-performing models for in vivo data from the Tractoinferno database. Overall, our results highlight both the potential and the limits of sequence-based deep learning models such as Transformers and RNNs for tractography, and emphasize the need for improved phantoms and evaluation methods for in vivo validation. We provide takeaways and recommendations for future researchers training and validating sequence-based supervised methods for tractography.
Machine Learning (cs.LG); Image and Video Processing (eess.IV); Neurons and Cognition (q-bio.NC)
arXiv:2610.01894 [cs.LG]
(or arXiv:2610.01894v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2610.01894 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 304 words
Related coverage
NVIDIA’s NV-Reason-CT brings AI reasoning to 3D CT scans
Quantum Zeitgeist · Sep 24, 2026
Quantum Wave Functions Emerge From Infinite Square Well
Quantum Zeitgeist · Jul 24, 2026
From Doyle to AGM: A Survey and an Implementation Roadmap for Belief Change
arXiv · Aug 18, 2026
Introducing NV-Reason-CT Open 3D CT VLM for Radiologist Chain-of-Thought Reasoning
NVIDIA · Sep 23, 2026