Research Projects

The project aims to explore the use of Electroencephalogram (EEG) and Functional Near Infrared Spectroscopy (fNIRS) derived signatures for different types of Concentrative and Mindfulness meditation practices. The study will employ Artificial Intelligence (AI)-based Machine Learning (ML) techniques to evaluate and validate the relevance of these signatures. At least 30 experienced meditators and 40 controls will be registered for the study, with participants selected through purposive sampling. EEG data will be collected using AD instruments' 16 channel data acquisition system and FNIRS data from BIOPAC systems' FNIR model 2000. EEG feature extraction will involve various time and frequency domain features, such as simple statistics, Hjorth features, fractal dimension of EEG signal, principal component analysis, independent component analysis, autoregressive coefficients, local discriminant bases LDB, shift invariant LDB, discrete wavelet transform DWT coefficients, coherence biomarker of EEG, and measures like oxyhemoglobin HbO, de-oxyhemoglobin HbR, and total hemoglobin. Statistical analysis will be conducted using SPSS174 software to assess the significance of extracted features in categorizing different meditative groups and controls. Traditional filter approaches will be implemented and evaluated, while new techniques based on evolutionary algorithms will be developed. Correlation analysis will be conducted to correlate the extracted signatures of State and Trait effects of different meditation practices across subjective experiences in meditators. AI-based machine learning algorithms will be developed for validation of derived signatures, including both offline and online systems. Automated machine learning models, advanced techniques like extreme learning and deep learning will be implemented and evaluated. Novel hybrid models based on multi-classifier concepts and majority voting will also be developed.