Independent component of 252-channel EEG data with bilateral occipital scalp projection. Derived by independent component analysis of approximately 700,000 points of data using runica() in the EEGLAB Toolbox, implementing extended infomax ICA. Here, the 252-channel data were reduced to 160 independent components by PCA prior to ICA training. Channels located below the head center are shown in this topoplot() as extending out from the model head borders. Note that ICA finds the gradient of the projection from the source to the skin surface not only at the scalp electrodes but also for channels on the neck, forehead and temples. Residual variance of the bet-fiting symmetric dual-dipole model in a model sphere head was 2.0%. Dipole calculations used the DIPFIT plug in.