Tensor clustering on outer-product of coefficient and component matrices of independent component analysis for reliable functional magnetic resonance imaging data decomposition

Guoqiang Hu,Qing Zhang,Abigail B Waters,Huanjie Li,Chi Zhang,Jianlin Wu,Fengyu Cong,Lisa D Nickerson

doi:10.1016/j.jneumeth.2019.108359

Abstract

BackgroundStability of spatial components is frequently used as a post-hoc selection criteria for choosing the dimensionality of an independent component analysis (ICA) of functional magnetic resonance imaging (fMRI) data. Although the stability of the ICA temporal courses differs from that of spatial components, temporal stability has not been considered during dimensionality decisions. New methodThe current study aims to (1) develop an algorithm to incorporate temporal course stability into dimensionality selection and (2) test the impact of temporal course on the stability of the ICA decomposition of fMRI data via tensor clustering. Resting state fMRI data were analyzed with two popular ICA algorithms, InfomaxICA and FastICA, using our new method and results were compared with model order selection based on spatial or temporal criteria alone. ResultsHierarchical clustering indicated that the stability of the ICA decomposition incorporating spatiotemporal tensor information performed similarly when compared to current best practice. However, we found that component spatiotemporal stability and convergence of the model varied significantly with model order. Considering both may lead to methodological improvements for determining ICA model order. Selected components were also significantly associated with relevant behavioral variables.Comparison with Existing Method: The Kullback–Leibler information criterion algorithm suggests the optimal model order for group ICA is 40, compared to the proposed method with an optimal model order of 20. ConclusionThe current study sheds new light on the importance of temporal course variability in ICA of fMRI data.

Highlights

Data-driven approaches for neuroimaging data analysis are useful for researchers because they do not require an a-priori hypothesis
In order to address this gap in the literature, we propose a new method, Tensor Clustering, which will allow us to assess both temporal and spatial stability in the independent component analysis (ICA) decomposition
The average of intra-cluster similarities for the cluster results based on the component matrix, coefficient matrix and spatiotemporal tensors are shown in Fig. 5 for FastICA and Fig. 7 for InfomaxICA

Summary

Introduction

Data-driven approaches for neuroimaging data analysis are useful for researchers because they do not require an a-priori hypothesis. Independent Component Analysis (ICA), the most commonly used approach, separates a multivariate signal into independent non-Gaussian sub-components (Bell and Sejnowski, 1995), and identifies spatiotemporal patterns that reflect both signals of interest and artifacts This approach has been widely used to analyze multiple modalities of neuroimaging with either multiple timepoints in a temporal series or multi-subject data in a subject-series, including structural Magnetic Resonance Imaging (sMRI) (Chen et al, 2014), functional MRI (fMRI) Selected components were significantly associated with relevant behavioral variables

Objectives

Methods

Results

Conclusion