Canonical correlation analysis with neural network for inter subject variability realization of EEG data

The detection of brain condition under different subjects is utmost important and it is a challenging task. EEG signals are such data that need to carefully analyze when it consists of series of different subjects. This paper explores the application of canonical correlation analysis with artificial neural networks for EEG data sets with different subjects and reference. We demonstrate the network's capabilities on EEG data to determine their subject to subject dependency in terms of correlation and then compare its effectiveness with that of a sine-cosine reference signals.