Spatio-temporal accuracy of ICA for FMRI

  title={Spatio-temporal accuracy of ICA for FMRI},
  author={Christian F. Beckmann and J. Alison Noble and Stephen M. Smith},
As an alternative to the General Linear Model (GLM), Independent Component Analysis (ICA) has been shown to identify spatiotemeporal modes that appear to correspond to meaningful source processes in the data [1,2]. In the case of ICA, both the mixing matrix and the component maps are estimated from the data using iterative techniques. In this abstract, we compare the consistency and accuracy of these estimates to rcsu1t.s obtained on artificial FMRI data using the GLM.