Would like to see the schematic and layout of the frontend board and a discussion on different design choices and tradeoffs. I designed something similar years ago (for Electrooculography, but it's the same principle) and getting a clean signal was by far the most challenging and interesting part of the project [0].
I'm not familiar enough with EEG to know whether the author's results are good enough to use it as a BCI, but it seems like they got it to work (with dry electrodes nonetheless). That's a win!
The Github repo says they're using the ADS1299, which is basically an all-in-one solution from TI for biopotential measurements. It's pretty popular, with lots of other examples to reference.
I'd assume they used the standard 5k series resistor/4.7nF filter capacitor arrangement the device recommends.
I'm not familiar enough with EEG to know whether the author's results are good enough to use it as a BCI, but it seems like they got it to work (with dry electrodes nonetheless). That's a win!
https://fellerts.no/projects/eog.php