Automatic detection of mental states, whether related to cognition or emotions, has numerous potential applications across various domains, including healthcare, education, neuroscience, and robotics.
This project focuses on classifying mental states using EEG signals collected from the Muse headband, offering a cost-effective and versatile solution for mental state classification.
The methodology for this project involves the following steps:
Find datasets containing EEG signals labeled with different mental states.
Clean and pre-process the data to remove noise and normalize features.
Extract relevant features from the EEG signals that can be used to distinguish between different mental states. These features may include statistical measures, frequency domain features, and phase synchronization.
Select the most important features to improve model performance and reduce overfitting. This can be done using various techniques like variance, correlation, mutual information, and ROC AUC.
Evaluate different machine learning models, such as Naive Bayes, Support Vector Machines (SVM), Decision Trees, Random Forests, and K-Nearest Neighbors (KNN), to classify mental states.
Train the best performing model on the entire dataset.
Test the trained model on unseen data to evaluate its generalizability.
- Data Acquisition: EEG recording device (e.g., Muse headband)
- Data Processing Libraries: MNE-Python, EEGLab
- Feature Extraction Libraries: EEGLab, SciPy, NumPy
- Feature Selection Libraries: scikit-learn
- Machine Learning Libraries: scikit-learn, TensorFlow
The Random Forest model emerged as the best performer, achieving an average accuracy of 97.4% using 10-fold cross-validation. This technique helps prevent overfitting and provides a more reliable estimate
