Performance of blind source separation algorithms for fMRI analysis using a group ICA method

Comparative Study

Nicolle Correa et al. Magn Reson Imaging. 2007 Jun.

Abstract

Independent component analysis (ICA) is a popular blind source separation technique that has proven to be promising for the analysis of functional magnetic resonance imaging (fMRI) data. A number of ICA approaches have been used for fMRI data analysis, and even more ICA algorithms exist; however, the impact of using different algorithms on the results is largely unexplored. In this paper, we study the performance of four major classes of algorithms for spatial ICA, namely, information maximization, maximization of non-Gaussianity, joint diagonalization of cross-cumulant matrices and second-order correlation-based methods, when they are applied to fMRI data from subjects performing a visuo-motor task. We use a group ICA method to study variability among different ICA algorithms, and we propose several analysis techniques to evaluate their performance. We compare how different ICA algorithms estimate activations in expected neuronal areas. The results demonstrate that the ICA algorithms using higher-order statistical information prove to be quite consistent for fMRI data analysis. Infomax, FastICA and joint approximate diagonalization of eigenmatrices (JADE) all yield reliable results, with each having its strengths in specific areas. Eigenvalue decomposition (EVD), an algorithm using second-order statistics, does not perform reliably for fMRI data. Additionally, for iterative ICA algorithms, it is important to investigate the variability of estimates from different runs. We test the consistency of the iterative algorithms Infomax and FastICA by running the algorithm a number of times with different initializations, and we note that they yield consistent results over these multiple runs. Our results greatly improve our confidence in the consistency of ICA for fMRI data analysis.

PubMed Disclaimer

Figures

Fig. 1

Components as estimated by SPM. Yellow indicates right task-related component and blue left task-related component.

Fig. 2

Components as estimated by a: Infomax, b: JADE, c: FastICA using tanh nonlinearity, d: EVD: The color code for the components is: Yellow: Right task related; Blue: Left task related; Red: Transiently task related; Orange: Temporal; Green: Default mode.

Fig. 3

Cluster plots for a: Infomax (cluster numbers 2, 3, 8, 10, and 16 correspond to the five components selected) and b: FastICA using gauss non-linearity (cluster numbers 1, 2, 4, 7, and 13 correspond to the five components selected), generated using ICASSO. Sin denotes the correlation between estimates within a cluster and sij denotes the correlation between estimates occurring in other clusters.

Fig. 4

The columns show the RTR, LTR and TTR cycle-averaged time courses (from left to right). The time courses in blue are of Infomax and those in red are of JADE in a, b, and c; of FastICA using tanh nonlinearity in d, e, and f; of FastICA using pow3 nonlinearity in g, h, and i; of FastICA using gauss nonlinearity j, k, and l; and of EVD in m, n, and o. The vertical lines indicate time points at which the time courses significantly differ.

Fig. 5

Clustering results for all algorithms for six clusters: Cluster 1: Temporal component as estimated by EVD: Cluster 2: Left task-related components; Cluster 3: Transiently task-related components; Cluster 4: Right task-related components; Cluster 5: Default components; Cluster 6: Temporal components.

References

1. Mckeown MJ, Makeig S, Brown GG, Jung T, Kindermann SS, Bell AJ, Sejnowski TJ. Analysis of fMRI by blind separation into independent spatial components. Hum. Brain Mapp. 1998;6:160–188. -PMC -PubMed
1. Bell AJ, Sejnowski TJ. An information maximization approach to blind separation and blind deconvolution. Neural Comput. 1995;7:1129–1159. -PubMed
1. Hyvärinen A, Oja E. A fast fixed-point algorithm for independent component analysis. Neural Comput. 1997;9:1483–1492.
1. Calhoun VD, Adali T, Pearlson GD. Independent component analysis applied to fMRI data: a generative model for validating results. J. VLSI Signal Process. Sys. 2004;37:281–291.
1. Esposito F, Formisano E, Seifritz E, Goebel R, Morrone R, Tedeschi G, Di Salle F. Spatial independent component analysis of functional MRI time-series: to what extent do results depend on the algorithm used? Hum. Brain Mapp. 2002;16:146–157. -PMC -PubMed

Performance of blind source separation algorithms for fMRI analysis using a group ICA method - PubMed (original) (raw)