Purchase this article with an account.
X. Zhang, J.C. Park, J. Hirsch, H.C. Donald; The Cortical Sources of the Multifocal VEP Revealed by ICA Analysis: Separating mfVEP Signal From Noise . Invest. Ophthalmol. Vis. Sci. 2006;47(13):1676.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The mfVEP offers a means to study local visual function and retinotopic organization with an electrophysiology measure [1,2]. However, a satisfactory means for removing the sources of EEG noise, such as an alpha wave, from mfVEP responses is still not available. ICA, which only relies on the response topology across the scalp, provides a means of isolating the alpha wave, which has the frequency characteristics of a VEP signal but a different topological distribution, from the mfVEP signal.
To understand the cortical sources of the multifocal visual evoked potential (mfVEP) signal and to separate mfVEP signal from EEG noise using multiple–electrode recording and Independent Component Analysis (ICA).
Monocular 60–sector mfVEPs were recorded for 14 mintues and for both the left and the right eyes with 30 electrodes. ICA analysis was performed to isolate the cortical sources of the mfVEP signal and noise. The signal to noise ratio (SNR) of the isolated mfVEP components were compared to the average SNR obtained with the standard 4–electrode bipolar placement.
ICA isolated only a few discrete sources for the120 responses of two eyes and rejected most components as noise. The components had very high SNR, retaining the waveform of the mfVEP waveform. The average SNR over 60 location responses for two subjects were 9.5 and 5.3, much higher than the SNR of 100 normal subject recorded with current 4 electrode method (mean 4.06).
ICA effectively isolates the mfVEP signal from noise. Although applying multiple electrodes adds considerable recording time, the improvement of the mfVEP SNR is significant. 1. Hood & Greenstein (2003). 2. Slotnick, et al.(1999).
This PDF is available to Subscribers Only