BrainDrain Using Machine Learning and Brain Waves to Detect Errors in Human
summary
Uses the Muse Headband to record EEG, cleans CSVs, aligns timestamps, and trains Python models in Keras; experiments use raw FFT features vs a recommended wavelet transform. Tasks are three‑digit multiplication and division with mostly male hackathon volunteers (many runs discarded for sensor movement), average session ~20 minutes. A simple LSTM/RNN barely outperforms decision trees and SVMs after long training (overfitting emerges near ~1,500 epochs). Prototype runs a Tkinter real‑time color feedback GUI.
AI-written summary · about
chapters (14)
- 0:00BrainDrain project intro
- 2:05Problem and goals
- 3:10Brain waves and EEG devices
- 5:15EEG seizure detection research
- 5:58Data collection and experiments
- 7:44Data visualization with R
- 13:50Correct vs incorrect solving
- 14:53Controlled math experiments
- 15:56Data cleaning and prep
- 17:49Early neural network models
- 19:39RNN vs traditional classifiers
- 20:21Keras LSTM update
- 22:08Real-time feedback prototype
- 24:15Next steps and future
related talks
31mLearn to control your brain: Brain Computer Interfacing with Python
35mTranslating Brain Electrical Signals into Actions through Models...
46mControlling apps with your mind
36mMachine-Brain Intelligence in Brain-Computer Interfaces
53mBrain: The ultimate hack
34mReading Thoughts with JavaScript
01h 01mKeynote: Using EEG and Machine Learning to Perform Lie Detection
31mJennifer Marsman, Principal Developer Evangelist, Microsoft