Application of Principal Component Analysis and Multivariate Signal Processing to Multichannel EEG and Neural Data
Abstract
Neuroscience (brain science) is changing at a very rapid pace. New brain-imaging technologies are allowing increasingly huge data sets, but analysing the resulting Big Data is one of the major struggles in modern neuroscience which is faced by all neuroscientists and neurological doctors. The increases in the number of simultaneously recorded data channels allows new discoveries about spatial and temporal structure in the brain, but also presents new challenges for data analyses. Since the EEG data are stored in matrices, algorithms developed in linear algebra are extremely useful.The paper discusses some matrix-based data analysis methods in neural time series data, with a focus on multivariate dimensionality reduction and source-separation methods. This includes covariance matrices, principal components analysis (PCA), generalized eigendecomposition and independent components analysis (ICA).