An improved multi-objective optimization-based CICA method with data-driver temporal reference for group fMRI data analysis

Abstract

Group independent component analysis (GICA) has been successfully applied to study multi-subject functional magnetic resonance imaging (fMRI) data, and the group independent component (GIC) represents the commonality of all subjects in the group. However, some studies show that the performance of GICA can be improved by incorporating a priori information, which is not always considered when looking for GICs in existing GICA methods. In this paper, we propose an improved multi-objective optimization-based constrained independent component analysis (CICA) method to take advantage of the temporal a priori information extracted from all subjects in the group by incorporating it into the computational process of GICA for group fMRI data analysis. The experimental results of simulated and real data show that the activated regions and the time course detected by the improved CICA method are more accurate in some sense. Moreover, the GIC computed by the improved CICA method has a higher correlation with the corresponding independent component of each subject in the group, which means that the improved CICA method with the temporal a priori information extracted from the group can better reflect the commonality of the subjects. These results demonstrate that the improved CICA method has its own advantages in fMRI data analysis.

Target Signal Extraction Method Based on Enhanced ICA with Reference

“…Firstly, we use the method of cICA to carry out simulation experiments [27]. In general, in cICA we do not know the correct threshold parameter ξ.…”

“…To solve the problem of setting threshold parameters ξ, some improved ICA-R algorithms are formulated [22,26,27].…”

“…Recently, Shi et al proposed a new model of ICA with reference signal (ICA-R), where an adaptive weighted summation method is introduced to solve the multiobjective optimization problem with a new fixed-point learning algorithm [26,27]. is method solves the threshold setting problem of cICA and effectively overcomes the problem of false extraction but faces the problem of determining the weight parameter.…”

Denoising brain networks using a fixed mathematical phase change in independent component analysis of magnitude‐only fMRI data

“…The second way is by averaging ICA spatial maps across multiple subjects (Calhoun, Adalı, Mcginty, et al, 2001 ; Kuang et al, 2018 ; Kuang, Lin, Gong, Chen, et al, 2017 ; Kuang, Lin, Gong, Cong, et al, 2017b ; Shi & Zeng, 2018 ). The third one is by detecting a shared brain network via the temporally concatenated group ICA (Britz et al, 2010 ; Calhoun et al, 2009 ; Calhoun & de Lacy, 2017 ; Erhardt et al, 2011 ; Qi et al, 2019 ; Qin et al, 2018 ; Shi et al, 2018 ), or via the tensor decomposition of multiple‐subject fMRI data (Acar et al, 2019 ; Beckmann & Smith, 2005 ; Han et al, 2022 ; Kuang et al, 2015 , 2020 ; Mørup et al, 2008 ; Wolf et al, 2010 ; Zhou et al, 2016 ). In these situations, amplitude‐based thresholding in terms of z ‐score is widely used for denoising ICA brain networks.…”

“…With the significantly increasing studies in ICA of magnitude‐only fMRI data, the z ‐score thresholds varied with different analyses, possibly due to different considerations such as for clearer visualization and closer to the reference. For example, z ‐score thresholds from 1 to 3 were predefined for denoising ICA spatial maps from a single subject (Brookes et al, 2011 ; Calhoun, Adalı, Pearlson, et al, 2001b ; Calhoun & de Lacy, 2017 ; Correa et al, 2005 ; Damoiseaux et al, 2007 ; Jung et al, 2001 ; Kuang, Lin, Gong, Cong, et al, 2017a ; Li et al, 2007 ; Long et al, 2009 ; Schwartz et al, 2019 ; Sui et al, 2012 ; Yu et al, 2015 ); z ‐score thresholds ranging from 0.5 to 2.5 were used for denoising an averaged ICA spatial map across multiple subjects (Calhoun, Adalı, Mcginty, et al, 2001 ; Kuang et al, 2018 ; Kuang, Lin, Gong, Chen, et al, 2017 ; Kuang, Lin, Gong, Cong, et al, 2017b ; Shi & Zeng, 2018 ; Yu et al, 2015 ); and z ‐score thresholds from 0.5 to 2.7 were used for denoising a shared spatial map obtained by the temporally concatenated group ICA (Britz et al, 2010 ; Calhoun et al, 2009 ; Calhoun & de Lacy, 2017 ; Erhardt et al, 2011 ; Qi et al, 2019 ; Qin et al, 2018 ; Shi et al, 2018 ) or by the tensor decomposition of multiple‐subject fMRI data (Acar et al, 2019 ; Han et al, 2022 ; Kuang et al, 2015 , 2020 ; Wolf et al, 2010 ).…”