Redesign: how does it all work?

Very cool experiment!

Not quite working, yet

There is still a lot of work to be done.

Some changes have been introduced to the active electrodes, trying to make them simpler, higher shielding, silver plated connectors, ...

As any good entry, here is a pic of the system actual status.

Next step will be redesign the hardware and perform some tests and simulations before mounting the whole system together. Old-school-baby-steps methodology has never let me down :)

Board is under construction ...

The board has its components already soldered. Short-circuit tests have been run and connections need to be added. Possibly the whole system in addition to the microcontroller would be boxed, not only to keep it neat and tight but safe from possible ESD (Electro-Static-Discharges), since I finally didn't include this block to simplify.

The milling machine

The LPKF ProtoMat® S42 introduces a new entry-level circuit board plotter for in-house rapid PCB prototyping. This compact system provides precision and performance for quickly and easily milling and drilling circuit board prototypes in a single day. In-house PCB prototyping eliminates production delays and the high cost of outside vendors, reducing a product's development time and time-to-market dramatically. Design data also remains securely in-house and under control. The S42 in particular is a perfect entry-level tool for educational and other settings where economy is a critical issue.

Reference: http://www.lpkfusa.com/protomat/s42.htm

Boards have been milled!

top side

bottom side

AE top and bot sides

Parts have arrived! :)

Circuit design completed v1.0

The first version of the board is completed and ready to etch!

Circuit schematic:

The schematic illustrates self explanatory dashed blocks and a name describing its main functionality. Together they form the whole structure of the "main" board. This acts as an interface between the microcontroller and AE (Active Electrodes, i.e. sensors).

PCB schematic:

First an image capture of the layout "as it is".

Secondly, we can distinguish the different sections that form the block-structure described above.

New version pending: shield-arduino compatible ...

Project Summary

Application idea and Objectives

The project’s main idea is the development of a sensor system, based on electrodes able to measure psychic patterns. The main patterns under study include variations in the occipital area where the primary visual cortex (V1) is. This area is responsible of the ionic fluctuations derived from the human visual system.

These alterations can be performed by the user at will by simply blinking the eyes. By identifying different patterns, we will be able to assign different functionalities to each of them, and hence create a human-computer interface.

Theory

            The EEG (Electroencephalogram) is normally described in terms of rhythmic activity divided into different frequency bands, ranging from 0 up to 100 Hz. Scientific research has demonstrated that the mayor differences when the eyes are open in contrast when the eyes are closed, are found in the “alpha” band (8 to 12Hz). Please notice the difference in amplitude in the following graphical representation of the output spectral analysis around the alpha band (around 20dB).

As the key denotes, the red curve represents the spectral activity when the eyes are open in contrast to the blue curve, that represents when the eyes are closed.

Positioning of electrodes is a question to be answered empirically in this project. Electrode locations are specified by the international “10-20 system” map shown on the second diagram below. Although neural activity is channeled over the occipital area (O1, O2) as seen from the first figure, different locations obtain different activity fluctuations, as seen on the figure below. It is a temporal representation of sensor output regarding different spatial positioning of the electrodes, when a single blink is performed.

Image from the following video:

As it can be seen, both areas show important fluctuations. Two electrodes should be enough to be able to distinguish between both states (eyes open, eyes closed) in O1, O2 locations or Fp1, Fp2 locations. Alternatively, four electrodes will clearly reinforce the information sampled.

Hardware Software – Block Diagram

From an electrical point of view, the measurement of brain activity is quite a challenge since the order of magnitude of these signals is around 10 to 100 microvolts (for measurements on the scalp surface for a typical adult human). The following block diagram describes different parts of the system.

It is basically formed by a chain of components: a pair of active dry electrodes “pick” the signal input that will be pre-amplified and introduced to a differential instrumentation amplifier; from here the signal is cleaned from common noise and amplified to be then introduced to an ADC; posteriorly digital data will be sent via a serial port communication to finally be treated by a mathematical software as Matlab.

A LPF may be added to reject noise frequencies above 100Hz, although this option is still under study. Note that active dry electrodes do not need conductive gel to work.