EEG Emotion Classification with Deep Learning

Overview

This repository contains Python code for performing emotion classification using EEG (Electroencephalogram) data. Emotion classification from EEG signals is an important application in neuroscience and human-computer interaction. The code leverages deep learning techniques to analyze EEG data and predict emotional states.

Key Features

1. Data Loading and Preprocessing:

   • Loads EEG data from a CSV file containing features extracted from EEG signals.
   • Converts emotion labels (e.g., 'NEGATIVE', 'NEUTRAL', 'POSITIVE') to numerical values (0, 1, 2) for classification.

2. Data Visualization:

   • Utilizes various visualization techniques to gain insights into the EEG data:
     • Pie chart to display the distribution of emotions in the dataset.
     • Time-series plot to visualize EEG signals over time.
     • Power Spectral Density (PSD) plot for spectral analysis.
     • Correlation heatmap to understand feature correlations.
     • t-SNE (t-Distributed Stochastic Neighbor Embedding) visualization for dimensionality reduction.

3. Feature Significance Analysis:

   • Conducts statistical tests (t-tests) to identify significant and non-significant features for each emotion category.
   • Presents results through bar charts, making it easy to understand which features contribute most to emotion prediction.

4. Advanced Preprocessing:

   • Normalizes feature data using z-score normalization.
   • Splits the dataset into training and testing sets.

5. Deep Learning Model:

   • Builds a deep neural network model for emotion classification.
   • The model architecture consists of multiple dense layers with ReLU activation functions and dropout layers.
   • Compiles the model using the Adam optimizer and sparse categorical cross-entropy loss.
   • Trains the model on the training data with validation, tracking accuracy during training.

6. Model Evaluation:

   • Evaluates the trained model on the testing dataset, reporting the accuracy of emotion classification.
   • Generates a confusion matrix and a classification report for a comprehensive performance assessment.

7. Random Sample Visualization:

   • Selects a random EEG sample from the testing dataset and predicts its emotion label.
   • Plots EEG signals from the selected sample for visualization.

Getting Started

For detailed usage instructions and explanations, please refer to the code comments and accompanying documentation in the repository.

Customization

Feel free to customize and extend this code for your specific EEG emotion classification projects.