Correlation analysis among single trials through frequency-domain EEG model for raw recordings and wavelet filtered MMN

Abstract

Event Abstract Back to Event Correlation analysis among single trials through frequency-domain EEG model for raw recordings and wavelet filtered MMN Fengyu Cong1*, T. Ristaniemi1 and H. Lyytinen2 1 Department of Mathematical Information Technology, University of Jyväskylä, Finland 2 Department of Psychology, University of Jyväskylä, Finland In the research of event-related potentials (ERPs), the single-trial study has become more and more attractive. This contribution is denoted to analyze the correlation among single trials through the frequency-domain EEG model for ERPs. It is interesting to compare such a correlation from the raw recordings and the wavelet filtered signal. Naturally, higher signal-to-noise ratio and better temporal structure of recordings should correspond to higher correlation. An auditory ERP, mismatch negativity (MMN), is taken as an example, but the method in this study is not limited to MMN. In the current experiment with 102 children MMN was elicited by two auditory deviants. Four children’s data were excluded due to too severe noise. With Electro-Cap International 20-electrode cap 350 trials of each deviant were recorded at C3,C4, Cz, F3, F4, Fz, Pz, M1 and M2 electrodes. Each trial last 650ms and the sampling rate was 200Hz. Usually, single-trial recordings are noisy and not well structured in ERPs experiments, moreover, ERP is time-locked, and hence these drive us to use the wavelet filter to remove noise and interferences. Reverse biorthogonal 6.8 wavelet and 7 levels were utilized for the decomposition, and coefficients of D4, D5, and D6 were adopted to reconstruct the desired signal under this wavelet. The raw recordings and wavelet filtered signals were transformed into the frequency-domain. We constructed a new complex-valued vector at each frequency bin, and the variable for the vector was the trial, then the frequency-domain EEG model was formulated. In our study this vector had 350 elements. As a result, the correlation between the real and imaginary parts reflects the correlation among trials. For each subject, the correlation coefficient was averaged over channels, deviants, and frequency bins within the optimal frequency band of MMN (2-8.5Hz). The grand averaged correlation coefficients over all subjects for the raw recordings and wavelet filtered signals were 0.178. and 0.365, respectively. The statistical tests among the two groups of correlation coefficients were F(1,97)=5966.0, P<0.000. This means that the difference was significant. In this study, the correlation among single trials was exploited in the frequency domain. In the raw recordings, the correlation was very small. After some noise and interferences were removed by the wavelet filter, the correlation became much higher. This implies that the quality of recordings has been improved by the wavelet filter. The correlation coefficient defined in this study would be a new parameter to qualify recordings in the ERPs experiment and data processing methods for EEG. Conference: MMN 09 Fifth Conference on Mismatch Negativity (MMN) and its Clinical and Scientific Applications, Budapest, Hungary, 4 Apr - 7 Apr, 2009. Presentation Type: Poster Presentation Topic: Poster Presentations Citation: Cong F, Ristaniemi T and Lyytinen H (2009). Correlation analysis among single trials through frequency-domain EEG model for raw recordings and wavelet filtered MMN. Conference Abstract: MMN 09 Fifth Conference on Mismatch Negativity (MMN) and its Clinical and Scientific Applications. doi: 10.3389/conf.neuro.09.2009.05.093 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 25 Mar 2009; Published Online: 25 Mar 2009. * Correspondence: Fengyu Cong, Department of Mathematical Information Technology, University of Jyväskylä, Jyväskylä, Finland, fengyu.cong@aliyun.com Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Fengyu Cong T. Ristaniemi H. Lyytinen Google Fengyu Cong T. Ristaniemi H. Lyytinen Google Scholar Fengyu Cong T. Ristaniemi H. Lyytinen PubMed Fengyu Cong T. Ristaniemi H. Lyytinen Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.