arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2606.09893eess.IVcs.AIcs.LG
收藏

TractFM:纤维束图基础模型

Tractogram foundation model

The State Key Lab of Brain-Machine Intelligence, Zhejiang University, Hangzhou(脑机智能国家重点实验室,浙江大学,杭州);Department of Radiology, Brigham and Women’s Hospital, Mass General Brigham, Boston(放射科,布里洛妇女医院,马萨诸塞总医院,波士顿);Harvard Medical School, Boston(哈佛医学院,波士顿);Academy of Medical Engineering and Translational Medicine, Tianjin University, Tianjin(医学工程与转化医学研究院,天津大学,天津);School of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu(信息与通信工程学院,电子科技大学,成都);Psychiatry Neuroimaging Laboratory, Brigham and Women’s Hospital, Mass General Brigham, Boston(精神病神经影像实验室,布里洛妇女医院,马萨诸塞总医院,波士顿);Department of Psychiatry, Center for Morphometric Analysis, Massachusetts General Hospital, Boston(精神病科,形态分析中心,马萨诸塞总医院,波士顿)

Guikun Chen, Yuqian Chen, Yijie Li, Yogesh Rathi, Nikos Makris, Fan Zhang, Wenguan Wang, Lauren J. O'Donnell

AI总结:

提出TractFM基础模型,直接从全脑纤维束集学习可复用表示,结合局部纤维编码器和置换等变纤维束编码器,通过密集解剖束分割预训练,实现纤维束级和受试者级任务的迁移。

摘要:

扩散MRI(dMRI)纤维束成像是在活体人脑中绘制白质通路的唯一非侵入性方法。它将每个大脑表示为一个纤维束图:一个大型、无序的三维流线集合,包含局部流线几何和全脑解剖组织的信息。这种结构使纤维束图成为表示学习的自然但具有挑战性的目标。现有方法将流线分类和受试者级预测视为独立问题:流线分类器关注几何模式,而受试者级预测通常依赖于手工特征。因此,当前方法无法学习连接流线解剖与全脑受试者间变异的可复用表示。本文介绍TractFM,一个纤维束图基础模型,直接从全脑纤维束集学习可复用表示。TractFM结合了局部流线编码器和置换等变纤维束编码器,使得一个受试者的所有流线能够在单次前向传递中共同上下文化。在密集解剖束分割(即给单个流线分配解剖标签)上的预训练产生了两种互补表示:用于束分割的上下文化流线级嵌入和用于下游受试者表型预测的紧凑受试者级描述符。在三种纤维束成像算法和五个dMRI数据集上,TractFM迁移到流线级和受试者级任务。其冻结表示实现了准确的束分割,并在独立数据集上预测年龄和性别。这些结果表明,全脑几何上下文(一次性学习)可以泛化到纤维束成像流程、数据集和预测任务中。

英文摘要:

Diffusion MRI (dMRI) tractography is the only noninvasive approach for mapping white-matter pathways in the living human brain. It represents each brain as a tractogram: a large, unordered set of three-dimensional streamlines that includes information about both local streamline geometry and whole-brain anatomical organization. This structure makes tractograms a natural but challenging target for representation learning. Existing methods treat streamline classification and subject-level prediction as separate problems: streamline classifiers focus on geometric patterns, whereas subject-level prediction often depends on hand-crafted features. As a result, current methods do not learn reusable representations that connect streamline anatomy with whole-brain inter-subject variation. Here we introduce TractFM, a tractogram foundation model that learns reusable representations directly from whole-brain streamline sets. TractFM combines a local streamline encoder with a permutation-equivariant tractogram encoder, allowing all streamlines from a subject to be contextualized jointly in a single forward pass. Pretraining on dense anatomical tract parcellation, i.e., assigning anatomical labels to individual streamlines, yields two complementary representations: contextualized streamline-level embeddings for tract parcellation and compact subject-level descriptors for downstream prediction of subject phenotypes. Across three tractography algorithms and five dMRI datasets, TractFM transfers to both streamline-level and subject-level tasks. Its frozen representations achieve accurate tract parcellation and predict age and sex across independent datasets. These results show that whole-brain geometric context, learned once, can generalize across tractography pipelines, datasets, and prediction tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑