NeuroAssist: Enhancing Cognition Computer Synergy with Adaptive AI & Enhance Neural Decoding
summary
Project builds a real-time BCI pipeline that classifies motor imagery from scalp EEG using reinforcement learning with DQ networks integrated with 1D CNN + LSTM and common spatial pattern feature extraction, bandpass filtering and Welch PSD. Training uses a 25-subject 60‑channel Giga dataset sampled at ~2500 Hz with channels C5 C3 C1 CZ C2 C4 C6 and reports up to 100% accuracy on Giga dataset and high accuracy on a separate 9‑subject 22‑channel BCI set. System maps predicted class to prosthetic commands and claims improved adaptability over traditional methods, with plans for broader testing.
AI-written summary · about
related talks
36mMachine-Brain Intelligence in Brain-Computer Interfaces
42mUnveiling the mysteries of Brain Computer Interfaces
18mAdding Human Learning in Brain--Computer Interfaces (BCIs): Towards a Practical Control Modality
11mConnected brains
05mBrain Computer Interface: Reading Human Brains Using Deep Learning
13mEnhanced EEG Data Classification with 3D SNN and On-Line Learning ESN
31mLearn to control your brain: Brain Computer Interfacing with Python
53mBrain: The ultimate hack