• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Asilomar Conference on Signals, Systems and Computers >Robust tensor decomposition of resting brain networks in stereotactic EEG
【24h】

Robust tensor decomposition of resting brain networks in stereotactic EEG

机译：立体定向脑电图中静态大脑网络的稳健张量分解

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Stereotactically implanted Electro-Encephalography (SEEG) in patients with epilepsy provides a unique insight into spontaneous human brain activity. Exploring dynamic functional connectivity in spontaneous SEEG signals provides a rich framework for studying brain networks. Tensor decomposition is a powerful tool for decoding dynamic networks, capturing the intrinsic interactions between multiple dimensions with less restrictive constraints than traditional 2D matrix decomposition methods such as PCA and ICA. Tensor decomposition, however, is seldom used for decoding large resting brain datasets due to its high computational complexity and poor robustness. In this paper, we describe a Scalable and Robust Sequential Canonical Polyadic Decomposition (SRSCPD) framework that can sequentially and robustly identify tensor models of successively higher rank. We demonstrate that SRSCPD is not only more robust than the popular Alternating Least Square (ALS) algorithm, but can also be extended to large-scale problems.
机译：立体定向植入的脑电图（SEEG）在癫痫患者中提供了对自发性人脑活动的独特见解。探索自发性SEEG信号中的动态功能连接性为研究大脑网络提供了一个丰富的框架。 Tensor分解是一种用于解码动态网络的强大工具，它可以以比传统的2D矩阵分解方法（例如PCA和ICA）更少的约束来捕获多维之间的固有交互。然而，由于张量分解的高计算复杂性和较差的鲁棒性，很少将其用于解码大型静息大脑数据集。在本文中，我们描述了一种可扩展且健壮的顺序正则多态分解（SRSCPD）框架，该框架可以顺序且可靠地识别相继较高秩的张量模型。我们证明，SRSCPD不仅比流行的交替最小二乘（ALS）算法更强大，而且还可以扩展到大规模问题。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号