Enhancing Meditation Techniques and Insights Using Feature Analysis of Electroencephalography (EEG)

Using a Muse 2 EEG headset connected via Bluetooth to a meditation app, this work describes a method for collecting, preprocessing, and classifying EEG data with IoT-based techniques. After preprocessing (removing redundant columns, handling missing data, normalizing, filtering), statistical features (mean, standard deviation, entropy) were extracted from the EEG signals. Three machine-learning models were trained: Support Vector Machine (SVM), Random Forest, and Multi-Layer Perceptron (MLP). The Random Forest model achieved the highest accuracy at 0.999, followed by MLP at 0.99 and SVM at 0.959. The authors propose this IoT-integrated approach as a viable method for brain-computer interfaces and meditation practice.

Study at a glance

Abstract

Through a Bluetooth connection between the Muse 2 device and the meditation app, leveraging IoT capabilities. The methodology encompasses data collection, preprocessing, feature extraction, and model training, all while utilizing Internet of Things (IoT) functionalities. The Muse 2 device records EEG data from multiple electrodes, which is then processed and analyzed within a mobile meditation platform. Preprocessing steps involve eliminating redundant columns, handling missing data, normalizing, and filtering, making use of IoT-enabled techniques. Feature extraction is carried out on EEG signals, utilizing statistical measures such as mean, standard deviation, and entropy. Three different models, including Support Vector Machine (SVM), Random Forest, and Multi-Layer Perceptron (MLP), are trained using the preprocessed data, incorporating Internet of Things (IoT) based methodologies. Model performance is assessed using metrics like accuracy, precision, recall, and F1-score, highlighting the effectiveness of IoT-driven techniques. Notably, the MLP and Random Forest models demonstrate remarkable accuracy and precision, underlining the potential of this IoT-integrated approach. Specifically, the three models achieved high accuracies, with Random Forest leading at 0.999, followed by SVM at 0.959 and MLP at 0.99. This study not only contributes to the field of brain-computer interfaces and assistive technologies but also showcases a viable method to seamlessly integrate the Muse 2 device into meditation practices, promoting self-awareness and mindfulness with the added power of IoT technology.