Home made EEG utilizing basic electronic components and an Arduino.
This project aims to developing an eletroencefalogram (EEG) utilizing simple eletronic and microeletronic components. The longer I get through my engineering course, the more I see how important is the "Cheapening Engineering" for the science development. So many projects that we found out so amazing and greatefull just takes really long or can't go any further because lack of investiment. Also, sometimes, several studies and tests does not require a very precise equipment, people usually look first for either prototypes or simpler versions of the device in question. Here, I'm trying to bring an alternative for those who looks to make some basic reseach in the area of brain computer interface (BCI), and also you can take that as an inspiration to build your own EEG or similar and even look for possible upgrades, fixes and so on.
If you either want to suggest changes, discuss or anything else, feel free to contact me: LinkedIn, or just use github issue if you prefer. Please, I'd love to hear your opnion!
The circuit is basically composed of an Arduino UNO as microcontroler, an amplifier, AgCl electrodes, an high and low pass filters, and a some multiplexers. I'll cover the use of them in more detail in the following subsections. To summarize, The filters will be used together as a band pass filter, so we get the output closest to only the frequencies emited by the brain. Those EEG's sign are of the order of microvolts, so we want to amplify them (of course with the amplifier) to get a better reading of 'em. The electrodes, attached to our head, will be used to collect the brain waves. In this case, we first need 3 electrodes. One will be used as ground, so we usually want to put at somewhere with no brain waves emission (like th ear), and the other two set what we call a channel. The more channels we have, the more measuresments we get, but we can't just connect the channels in series or parallel. For this to work, I found that using multiplexers could be a solution. They basically will constantly alternate between which channel the measures are made. Lastly, the Arduino is used for real time measures of this filtered and amplified sign and then we can do any needed analyses with any programming language. For this work, we're only focus on the arduino set up with C++ and python for the data analysis, but other could've be used.
The AMP used is an AD620, which can be powered by a single +5V source and a common ground point. This is an operational AMP, that means, it amplifies the difference between two input signals. We connect the MUX+ to the S+ pin and the MUX- to the S-. Both Vneg and GND pin is going to be connected to the circuit common ground, Vin is connected to Arduino's +5V and Vout will continue to the filters. The AD620 gain is adjusted with an external resistor, and is given by,
An BPF can be made with an High-pass Filter (HPF) togheter with a Low-pass Filter (LPF). An HPF removes frequencies below an certain limit, allowing just the highest frequences to get through. For EEG's device, the cutoff frequency is usually between 0.5 and 1Hz, to remove frequences related to moviments (for this first approach, I'm just interested in evaluating the device capacity/accuracy). On the other hand, the LPF removes the frequencies above an certain limit. For EEG's device, usually the cutoff frequency is arround 30 to 50Hz, to eliminate the high frequency noise, including eletrical grid interferencies (60Hz). The cutoff frequency can be calculated by,
The filters are made of an resistor and capacitor (for each filter) direct connected right in between the AMP ouput (Vout) and the Arduino. For a 0.5-50Hz range frequency and, for both filters, an 330k
The signal reading is made with the electrodes kit Neurosoft NS-DECL101500-PC. To have multiple channels measurements, it must use MUX's (or at least, that's the solution I end up with). First things first, we need 2 MUX, the "positive mux" (MUX+) is gonna hold the positive electrodes while the "negative mux" will get to work with the negative electrodes. To start, both must be connected to the +5V Arduino pin and the circuit's commom ground. The MUX+ SIG port is going to the S+ (or S- for MUX-) AMP pin. Since I'm using just 2 chanells, I went for C0 and C1 (first input channels) that both outputs to the S0 pin (you can also check an multiplexer tutorial, for more detail or better explanation, if wanted).
An schematic of this first version of the device can be seen next. It was made using the Software EASY EDA.
This project is licensed under the MIT license - See LICENSE for more details.