EEG Controlled Pong

Overview

Our project involved the design and implementation of a circuit that would be able to utilize EEG signals to control lights and play classic Pong

Description

Analog Front End (AFE)

• Low-noise analog front end that filters out unwanted noise and amplifies the signal by a gain of around 1500

• AFE Overview:

  1. AD623 Instrumentation Amplifier - Amplifies the difference between electrode inputs while removing environmental noise

  2. High-pass filter (referenced to ground) - Clears out low frequencies below 0.05Hz, removes DC offset, and re-centers the signal at a virtual ground of 1.65V

  3. MCP6001 Op-amp (inverting configuration) - Low impedance, doesn't affect AC outputs; introduces a gain of about 65 (default)

  4. Low-pass filter - Removes frequencies above about 35Hz

  5. Potentiometer - Modulates amplitude by adjusting feedback resistance to the AMP

  6. MCP3208 ADC - Converts AFE output to digital signals for MCU interpretation

MCU

• Connected over USB-C to the computer for performing math operations on the signal to make it 'readable', and converting them into discrete steps to be fed into the output programs. The signals that we were intending to use were P300 events (eye-blinks - http://psy.hull.ac.uk/Staff/m.large/Infonew/),  and possibly patterns in Beta waves (residing in 13Hz to 30Hz region, though this would require some iterative testing to extract usable events).

• Three programs to be run on MCU -

  1. A signal cleanup, plotting and extraction code.

  2. An initial program for controlling LED array to indicate successful interpretation of brain signals.

  3. A final program for classic Pong (displayed on the laptop in this version)

Final Cost of PCB

For one board, the final cost of the PCB and components came to $29.37.