Share Email Print

Proceedings Paper

Real-time independent component analysis of fMRI using spatial constraint
Author(s): Zhi Wang; Hang Zhang; Xia Wu; Li Yao; Zhiying Long
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Real-time functional magnetic resonance imaging (fMRI) is a useful tool that researchers can monitor and assess dynamic brain activity in real time and train individuals to actively control over their brain activation by using neurofeedback. Independent Component Analysis (ICA) is a data-driven method which can recover a set of independent sources from data without using any prior information. Since ICA was firstly proposed to be applied to fMRI data by Mckeown (1998), it has become more and more popular in offline fMRI data analysis. However, ICA was seldom used in real-time fMRI studies due to its large time cost. Although Esposito (2005) proposed a real-time ICA (rtICA) framework by combining FastICA with a sliding-window approach, it was only applied to analyze single-slice data rather than full-brain data and was not stable. The semi-blind rtICA (sb-rtICA) method proposed by Ma (2011) can reduce the computation time and improve stability by adding regularization of certain estimated time course using the experiment paradigm information to rtICA. However, the target independent component (IC) cannot be extracted as the first component in all sliding windows by sb-rtICA, which still adds computation time to some extent. The constrained ICA proposed by Lu (2005) can eliminate the ICA’s indeterminacy on permutation. In this study, we proposed a real-time Constrained Independent Component Analysis (rtCICA) method by combining CICA with the sliding-window technique to improve the performance of rtICA. The basic idea of rtCICA is to induce spatial prior information as constraints into ICA so that the target IC can be always automatically extracted as the first one. Both simulated and real-time fMRI experiments demonstrated that rtCICA outperforms rtICA greatly in the stability and the computational time.

Paper Details

Date Published: 29 March 2013
PDF: 10 pages
Proc. SPIE 8672, Medical Imaging 2013: Biomedical Applications in Molecular, Structural, and Functional Imaging, 86720P (29 March 2013); doi: 10.1117/12.2006512
Show Author Affiliations
Zhi Wang, Beijing Normal Univ. (China)
Hang Zhang, Peking Univ. (China)
Xia Wu, Beijing Normal Univ. (China)
Li Yao, Beijing Normal Univ. (China)
Zhiying Long, Beijing Normal Univ. (China)

Published in SPIE Proceedings Vol. 8672:
Medical Imaging 2013: Biomedical Applications in Molecular, Structural, and Functional Imaging
John B. Weaver; Robert C. Molthen, Editor(s)

© SPIE. Terms of Use
Back to Top