Open AccessAn Estimation Method for Multi‐channel EEG Data Based on Canonical Correlation Analysis
Author(s) -
Yan Zheng,
Wan Xiaojiao,
Ling Chaodong
Publication year - 2015
Publication title -
chinese journal of electronics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.267
H-Index - 25
eISSN - 2075-5597
pISSN - 1022-4653
DOI - 10.1049/cje.2015.07.022
Subject(s) - canonical correlation , computer science , electroencephalography , correlation , pattern recognition (psychology) , channel (broadcasting) , invariant (physics) , algorithm , artificial intelligence , noise (video) , data mining , mathematics , psychology , geometry , computer network , image (mathematics) , psychiatry , mathematical physics
Electroencephalogram (EEG) signal is oftencontaminated by electronic noise as well as movementartifacts. This paper presented an algorithm basedon Canonical correlation analysis (CCA) to estimate multichannelEEG data. Different from previous studies, inwhich CCA was mainly used to detect the invariant featuresspecific to each brain state, in this paper, the canonicalvariates computed by CCA were used to reconstructthe multi‐channel EEG data. Firstly, two data sets, EEGsignals and the reference signals based on prior knowledgewere constructed. Next, canonical variates were computedby projecting the two data sets onto basis vectors. Finally,a least squares solution was used to estimate the multichannelEEG data. The experiment results suggested thatthe algorithm is capable of reconstructing the actual specificcomponents with high quality. We also hint futurepossible application of the algorithm in the estimation offunctional connectivity patterns at the end of the paper.