Single Channel Analysis of Electromagnetic Brain Signals Through ICA in a Dynamical Systems Framework

This paper introduces a method for extracting information from single channel recordings of electromagnetic (EM) brain signals. In a dynamical embedding framework, the measured electroencephalogram (EEC) and magnetoencephalogram (MEC) signals are assumed generated by the non-linear interaction of a few degrees of freedom. In a three-step process, first an appropriate embedding matrix is constructed out of a series of delay vectors from the measured signal. Then independent component analysis (ICA) is performed on the embedding matrix to decompose the single channel recording into its underlying independent components (ICs). The ICs are treated as a convenient expansion basis and subjective methods are then used to identify components of interest relevant to the application. These ICs are then projected back onto the measurement space in isolation. The method has been applied to single channels of both EEC and MEC recordings and is shown to isolate, amongst others: i) artifactual components such as ocular, electrocardiographic and electrode artifact, ii) seizure components in epileptic EEC recordings and iii) theta band, tumour related, activity in MEC recordings.
Papers from 23rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society, October 25-28, 2001, held in Istanbul, Turkey. See also ADM001351 for entire conference on cd-rom.