Graphical Model-Based ICA Applied to the Analysis of fMRI and MRI Time Series
Anke Meyer‐Baese, Axel Saalbach · IEEE International Conference on Neural Networks/IEEE ... International Conference on Neural Networks · 2007
Graphical model-based independent component analysis (ICA) represents a novel and powerful paradigm in exploratory data analysis for biomedical imaging applications. Two very important techniques, tree-dependent and topographic ICA, implement a clustering of dependent components by demixing and classifying time series of pixels exhibiting similar properties of local signal dynamics. The theoretical background is presented in the beginning, followed by several medical applications demonstrating the flexibility and conceptual power of these techniques. These applications range from functional MRI data analysis to breast MRI. For fMRI, these methods can be employed to identify and separate time courses of interest, along with their associated spatial patterns. In breast MRI, a detection of the lesion is achieved and in addition a subclassiflcation is obtained within the lesion with regard to regions characterized by different MRI signal time-courses. In the present paper, we conclude that graphical model-based ICA techniques provide a robust method for blind analysis of time series image data in the important and current field of functional and dynamic MRI.