Open AccessScaled correlation analysis of electroencephalography: a new measure of signal influence
Author(s) -
Ahirwal Mitul K.,
Kumar Anil,
Singh Girish K.,
Londhe Narendra D.,
Suri Jasjit S.
Publication year - 2016
Publication title -
iet science, measurement and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.418
H-Index - 49
eISSN - 1751-8830
pISSN - 1751-8822
DOI - 10.1049/iet-smt.2015.0299
Subject(s) - electroencephalography , measure (data warehouse) , signal (programming language) , correlation , linear correlation , signal processing , computer science , pattern recognition (psychology) , biological system , acoustics , mathematics , artificial intelligence , statistics , physics , data mining , telecommunications , psychology , radar , geometry , programming language , psychiatry , biology
Electroencephalography (EEG) signals recording are the mixture of electrical potentials generated from different sources. These signals are influenced by different potentials. Currently, there exists no measure that can evaluate the measure influence among the signals. A new measure of influence has been proposed based on the distribution of correlation (DCOR) that quantifies the relative influence of constituent sub‐band signals over full band signal. To estimate the inter‐influence, scaled correlation analysis of signal sub‐components is investigated. Results so obtained demonstrate that the signal influence of highest‐frequency components present in the signal is more in case of linear/stationary signals, compared with non‐linear/non‐stationary (EEG/event related potential) signals. These findings are concluded with two types of analysis: (i) mixed influence analysis and (ii) mutual influence analysis. It is demonstrated that for separation of negative and positive correlations (CORs) using the proposed novel measure of signal influence (DCOR) is 14.24% better than other conventional COR method.