Freitag, 22. März 2013

Building an EEG input device for BCI experiments

Lately I read a couple of interesting articles regarding brain computer interfaces or BCI. With these devices you can use your thoughts to control things. I decided to give this a try and build myself an EEG to start playing around. The EEG is needed for the data acquisition, the rest of the work is done in software.

I based my design on the openEEG layout. The thing I changed is the instrumental amplifier, I went with the LT1167, a high quality IC, which I sampled for free. I also changed the way the virtual ground signal is generated. I used an IC I had lying around, a MAX931 a comparator chip which happens to have a 2% accurate reference voltage on board. So I just used the chips internal ref voltage, which I amplified  to the needed value with an op-amp for the use as analog reference voltage, AnalogRef. I then divide AnalogRef by two with a simple resistor divider and buffer and low pass this signal again to get my final virtual ground (VGND).
The Virtual Ground Circuit

VGND is needed to measure positive and negative voltage swings with just a positive power supply. Here is the link.

One EEG channel connects two electrodes, the channel's signal is then the difference between these Inputs. I read a good article called "Instrumentation Amplifiers for bioelectric events: a design with a minimal number of part". It explains all design choices in great detail and I believe served as inspiration for the openEEG design.

My design has two channels thus 4 electrodes. This is the link to the image below.
One EEG input channel which connects to two electrodes, IN_A and IN_B
The DRL circuit is used to summ the CMR signals of all the eeg channels up, buffer that signal and connect it to the body of the monitored person. It's use should result in a better signal quality. You will need an op-amp wired as buffer with a 10K resistor at it's output per channel!
The DRL circuit
 For the prototype I use SMD components wherever possible, mounted on breadboard. It is quite easy to assemble and the result are still very tiny, without the extra hassle to make a PCB. A SMD component like a resistor or a cap takes 2 drills, one for each side!

I will post pictures of the prototype as soon as i take them!

Here are the links to the EAGLE files:
The layout
The board file, which is routed in 2.54mm grid, to transfer it easily to breadboard.

Keine Kommentare:

Kommentar veröffentlichen