Well, it’s first post in this blog. I’ll start from my current project – mobile EEG brain activity classifier, based on neural network. I made some calcutations of analog parts. Below you can see a circuit diagram.
Special attention should be given to the equivalent circuit of contact between the electrodes and the head. I used the basic circuit . My signal propagation model of brain (Fig. left) doesn’t take into account the reactive components.
The electrode attachment points are shown in Fig. rihgt. Electrodes 1,2 (brass, area ~ 1.5 cm2) are fixed on the forehead at a distance of 2-3 cm. The reference electrode is fixed as close to the occipital part of the skull as possible (the proposed site of installation). To calculate the resistances, 2 measurements were taken (Fig. below).
Assuming the resistance of the contacts to be equal (by cleaning the skin with alcohol), and solving a trivial equations system, it is easy to find the corresponding resistances.
Below are the frequency response and the FFT of 10 Hz sinusoidal signal at various points in the circuit. As can be seen from the FFT, the signal-to-noise ratio in the main working frequency band (1-100 Hz) at the input (~ 80 dB) and the output (~ 70 dB) is not too different.
Please, if you have any comments about the material, please contact me via the “Contact” section.
- Jiawei Xu, Low-power active electrode for Wearable EEG, Page 5.