k-means: a new generalized k-means clustering algorithm
Pattern Recognition Letters
Hi-index | 0.00 |
A new equivalent representation of the neural electric activities in the brain is presented here, which images the time courses of neural sources within several seconds by the methods of k-means cluster and correlation coefficient on the scalp EEG waveforms. The simulation results demonstrate the validity of this method: The correlation coefficients between time courses of simulation sources and those of estimated sources are higher than 0.974 for random noise with NSR≤0.2. The distances between the normalized locations of simulation sources and those of estimated sources are shorter than 0.108 for random noise with NSR≤0.2. The proposed approach has also been applied to a human study related to spatial selective attention. The results of real VEPs data show that three correlative sources can be located and time courses of those can describe the attention cognitive process correctly.