EEG Data Analyser and Simulator

AI-driven simulation and prediction of brain dynamics using EEG data

Overview

This project aims to develop an AI-powered system that constructs a personalized Digital Twin of the Brain using EEG (Electroencephalogram) data.

The system leverages deep learning and signal processing to analyze, predict, and simulate neurological patterns — enabling early detection of epilepsy, cognitive stress, depression, and other brain-related conditions. The system also aims to classify emotions in awake and dream phase. Other implications include the prediction of failure and success of surgery and drug identification and simulation of its effects on EEG data.

In addition, the project explores neural generative modeling to interpret REM-phase brain activity into abstract visual representations, pushing the boundaries of dream analysis and subconscious understanding.

Objectives

• Develop a personalized digital twin of the brain using EEG data.
• Detect early neurological conditions such as epilepsy, stress and depression.
• Simulate effects of medications or surgical interventions.
• Build interpretable and clinically relevant models for neurological decision support.
• Explore deep generative modeling for dream visualization using REM EEG patterns.

Key implementations

• Epilepsy predictor :- Transformer based model to detect preictal/interictal EEG patterns to predict seizure.
• Stress classification :- Transformer based model to classify stress states like relaxed and stressed using EEG data.
• Surgery Outcome Prediction:- Transformer based model to predict success or failure of surgical intervention on Drug-resistant epileptic patients.
• Emotion Recognition:- LSTM based model to classify EEG signals into emotional states (negative, neutral, positive).
• Dream Emotion Prediction:- LSTM based model to analyze REM EEG activity to infer underlying dream emotions.

Each implementation includes:

• Custom EEG preprocessing pipeline
• Feature extraction
• Deep learning model training and evaluation
• Visualization and interpretation of brain activity

Future work

• Personalized brain twin simulation
• Dream visualization via generative models (VAE/GANs)

Documentation

Documentation link : Documentation