Analysis of EEG Signals and Biomedical Changes Due to Meditation on Brain by Using ICA for Feature Extraction

Meditation imparts relaxation to the human being. Nowadays due to worrying and tense atmosphere, the meditation has become demanding solution to reduce physical and mental stress. The current experimentation aims at investigating the biomedical changes of meditation on the human brain. The term meditation refers to an extremely wide variety of practices. The central goal of these meditation practices is to alter mental states. However, the manner in which this goal is pursued in the variety of practices differs from person to person each so that as of today no clear definition of necessary and sufficient criteria for meditation exists that has achieved universal or widespread acceptance. Meditation can be characterized by concentration of attention on the breath, the recitation of a mantra, the state of calmness, mental freedom and the cultivation of a compassionate state. The major focus of fundamental research on meditation is to observe and analyze the changes in brain activation, functioning and anatomy induced by mental training. In this research work, 11 subjects with Mindfulness meditation practice and 11 subjects of non-meditation group are considered. The EEG signals are obtained by using 14 electrode systems. The pre-processing technique used is Daubechies wavelet transform before calculating the statistical features. The statistical features calculated are variance, kurtosis, energy and entropy. Along with these four features, the three features are calculated by using Fast Independent Component Analysis (fast ICA). These features are averaged and plot in order to categorize between meditational and non-meditational group.