Abstract
Constrained independent component analysis (CICA) is capable of eliminating the order ambiguity that is found in the standard ICA and extracting the desired independent components by incorporating prior information into the ICA contrast function. However, the current CICA method produces constraints that are based on only one type of prior information (temporal/spatial), which may increase the dependency of CICA on the accuracy of the prior information. To improve the robustness of CICA and to reduce the impact of the accuracy of prior information on CICA, we proposed a temporally and spatially constrained ICA (TSCICA) method that incorporated two types of prior information, both temporal and spatial, as constraints in the ICA. The proposed approach was tested using simulated fMRI data and was applied to a real fMRI experiment using 13 subjects who performed a movement task. Additionally, the performance of TSCICA was compared with the ICA method, the temporally CICA (TCICA) method and the spatially CICA (SCICA) method. The results from the simulation and from the real fMRI data demonstrated that TSCICA outperformed TCICA, SCICA and ICA in terms of robustness to noise. Moreover, the TSCICA method displayed better robustness to prior temporal/spatial information than the TCICA/SCICA method.
Highlights
Independent component analysis (ICA) is a data-driven method that can recover a set of maximally independent sources from observed multivariate data without using any prior information [1,2,3]
In contrast to the complementary univariate general linear model (GLM) method, which is performed on a voxel-by-voxel basis [8], the ICA method is able to extract multiple brain networks that are engaged in various elements of cognitive processing without any prior knowledge
The dimension of each simulated dataset were reduced by principal component analysis (PCA), with 99.9% of the total variance of the mixed signals retained before temporally and spatially constrained ICA (TSCICA), temporally CICA (TCICA), spatially CICA (SCICA) and ICA, to ensure that all the informative components were included [16]
Summary
Independent component analysis (ICA) is a data-driven method that can recover a set of maximally independent sources from observed multivariate data without using any prior information [1,2,3]. In contrast to the complementary univariate general linear model (GLM) method, which is performed on a voxel-by-voxel basis [8], the ICA method is able to extract multiple brain networks that are engaged in various elements of cognitive processing without any prior knowledge. This ability makes ICA an increasingly attractive exploratory tool to study functional brain networks either at rest [9] or during a cognitive task [10]. To improve the convergence of CICA, learning-rate-free CICA algorithms were proposed by Wang (2011) and were applied to separate spatially independent components from fMRI data using temporal constraints on the mixing matrix [17]. A priori information that was available in the spatial/temporal domain was fed to the first CICA stage, and the output of the first CICA stage was added to the second CICA stage as the constraint
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call