uHD EEG Machine Learning Models: A Spectacular Repository

Welcome to the uHD EEG Machine Learning Models Repository, where we aim to train and evaluate state-of-the-art machine learning models on Ultra-High Density Electroencephalography (uHD EEG) data to decode individual finger movements.

Dataset & Inspiration

The dataset used in this repository is available through the following URL:
Dataset

This incredible repository is inspired by the groundbreaking research article:
Individual Finger Movement Decoding using a novel Ultra-High Density EEG-based BCI system

Unveil the Secrets of the Dataset

Begin your journey with the inspect_data.ipynb script. It will guide you through the basics of the dataset and provide essential insights.

Feature Extraction: Unlock the Potential

The key to successful machine learning lies in feature extraction. The extract_features.ipynb script will extract features for a single subject and store the data in a convenient h5 file.

Training Binary Finger Classifications: Power at Your Fingertips

Support Vector Machines (SVM) Models

• Conquer SVM models for each finger pair for a single subject:
  train_binary_SVM.ipynb
• Master SVM models for each subject and store results in CSV:
  train_binary_SVM.py

Deep Learning Models: Dive into the Depths

• Train Deep Learning models for each finger pair for a single subject:
  train_binary_DL.ipynb
• Harness the power of DL models for each subject and store results in CSV:
  train_binary_DL.py

Make plots

• Process output CSVs and save visually stunning plots to file:
  visualize_results.ipynb

Saliency Plot: Visualize the Importance

Create breathtaking saliency plots based on the backpropagated value of a neural network:
saliency_plot_2classes.py

Training Classification for Single Subjects: Five Classes of Excellence

• train_DL_singleSubject_5classes.ipynb
• train_DL_singleSubject_5classes_RNN.ipynb

Training Neural Network on All Subjects: The Ultimate Power

• train_DL_all_subject.ipynb
• train_DL_all_subject_individual_first_layer.ipynb
• tain_DL_all_subject_individual_cross_attention.ipynb