Multi-feature extraction, analysis, and classification for control and meditators' electroencephalogram

Meditation has a metaphysical impact on human brain functioning. It is of utmost required to infer the cognitive effects of meditation using an electroencephalogram (EEG). In this novel work, the analyses of EEG signals' features are extracted for cognitive effects on a human brain for meditation intervention of 25 subjects. To analyze the meditation effects, this study examines the feasibility of statistical, spatial, spectral, coherence features, and time-frequency analysis of EEG signals for control and meditator group. Based on the effective features the various classifiers are used to compare the accuracy and distinguish a subject as control or meditator. The results demonstrate that the Support Vector Machine (SVM) gives better accuracy than Artificial Neural Network (ANN) and k-Nearest Neighbors (KNN). The statistical analysis shows that the Variance and Sample Entropy decreased in meditators whereas, in spatial analysis, the Mahalanobis distance increased. The spectral analysis stated that theta power has increased 88% of subjects whereas the alpha power is increased for the entire subjects after meditation. The coherence observed in the pre-frontal lobes' electrode pair is more in the meditators than in the control group. Eventually, meditation improves relaxation, cognitive functions, calmness, and mental concentration.

Automated summary

This study analyzed the cognitive effects of meditation on the human brain using electroencephalogram (EEG) signals. The results demonstrate that meditation improves relaxation, cognitive functions, calmness, and mental concentration. Statistical, spatial, and spectral analyses support these findings.