Status of research on brainwave reading
After much research, We came to the conclusion that it is not possible to obtain reliable information "watching" the brain from the outside.
It's not about budget, It's about seeing the brain from the outside, then only see the result of the sum of the activities of all neurons of the brain together (with minor differences due to the position of the electrodes – so minimal that much of the time it's just noise or interference from the movement)
Whereby "patient" cannot "think of a pressed key" but I get with a different activity. And press is just a side effect of this intense activity. Basically you must learn to involve the entire brain and sync, in order to increase, for example,, alpha waves compared to Beta. An ordinary person cannot do it right, the best they can do it in a few seconds, but just focusing and standing firm.
A bit like saying: "to press up arrow key-on you must stop breathing until you see the button is pressed, to release it again to breathe and after a few seconds you will release "
Perform actions with the thought ?
To perform actions there are better methods. If budget issue Hawking students (https://it.wikipedia.org/wiki/Stephen_Hawking) they equipped its super processed helmet Chair. But they tried everything and in the end they chose to look eye-movements with a camera.
With a camera, for example,, eye closure you discover very quickly. With the closing of the eyes you can get four bits of information and many more with eye movements, myoelectric impulses, the direction and speed of the breath etc… With these techniques can easily get tens of bits per second while with an EEG, for how expensive, the response times are of many seconds for a single bit.
For Hawking have done just as well, If you look carefully, in all its recent pictures you see a camera hanging eyeglasses (and in some cases even more experimental sensors).
And now Hawking can speak, Surf the internet and even write books. With the thought would never be able to do all these things.
If there was any possibility of using the thought wouldn't have budget problems. The only wheelchair costs as hundreds of MindWave and EmotivEEG and already had lithium batteries when none of us had even heard of.
Can we do better than the MindWave and EmotivEEG ?
Hate to say it, but we can't. Those two devices have been designed with great care and "trying" to obtain the best possible result. Yet the results are more or less confused (When it's good), and completely random (When the subject is not suitable for).
Subject unsuited? Apologies are always the same, you have to train more, It takes more concentration… In practice to press a button you have to relax completely and then focus for a few seconds. What you should do to write a message? Probably ten years of apprenticeship as a Buddhist monk.
Now the electrodes of a certain quality are all non-contact, but the results are always more or less the same: great instability and no security that if worked today will work tomorrow. Even with 24 endings and big devices with a closet have managed to have decent results.
Active electrodes (first version)
The first version, as from previous images, required a particular electrode with Silver tips. We thank Marco Vitali of: www.vitalibruno.com for having prepared a test enclosure in silver (see image). To make contact with the skin can do no better than silver, even gold is better, and toes help to go hair and to firmly hold the electrode.
Unfortunately, testing has shown that even with these electrodes just the slightest movement to create potential differences of the same order of magnitude of EEG signals (100 UV). To avoid this you have to degrease the skin with alcohol and press much with belt. And unfortunately it also happens that after some time the tips, as far as rounded, begin to annoy. In less than one hour you create markers in the skin which then lasts for a long time.
In conclusion, hate to say it, but this road is not reliable enough to be part of the system Theremino.
Active electrodes (non-contact)
Fortunately during rehearsals with the shell shield and silver with high impedance electrodes we discovered that even with a thin insulating layer between electrode and skin useful signals continued to pass by capacity.
We discovered the non-contact detecting EEG and ECG!!!
As usual, however it is difficult to get there first.
Just search “Do not contact electrodes” on Google…
Non-contact electrodes have many advantages, first of all the complete elimination of the galvanic connection with patient, then security to 100% without batteries or complicated power circuits with safety regulations impossible.
Advantages of the electrodes without contact:
- Total security (the patient is electrically isolated)
- Elimination of complicated isolation circuits
- Eliminate the need to go to batteries
- No galvanic contact voltages for which serves the silver
- Being able to use the copper you do everything with multiple layers of printed circuit board
- Better signal even if the patient moves
- “Zero preparation” (You should clean your skin with alcohol)
- No contact tips
- The flat surface and easier to sterilize
- Functioning even through clothing
- Reduced performance you get even a few centimeters away
- Reduced performance you can make fabric electrodes (moistened only by sweat)
First tests of the electrodes with signal generator
Those wishing to experiment with ThereminoEEG version 1.0 Please download it from here:
To test it you can use a pin set as input proportional Cap16 and move with a finger. If you move fast you can get beyond the 10 Hertz.
Or sooner or later we will start these new tests PCB. The center circle on the right is the electrode that mates with the skin, the ring (and the back face all in copper that is not seen) isolate it from the environment with an active screen that reduces the effective ability to practically zero.
For the first tests we chose an external diameter of 30 mm and the total thickness is below the 5 mm. Are electrodes small enough to fit comfortably in their place them under an elastic band, but at the same time have very 6 cm paintings of active electrode.
The round shape is a major improvement, it becomes much easier to build them mechanically, and then the containers.
Version 2 – compensation capacity
This new version has a more complex compensation circuit, According to the simulations, should improve the functioning even electrodes away from skin. With strong ECG signals you can work even at incredible distances from the skin (10 mm and up to 20 mm with some degradation of bandwidth)
We are doing experiments to determine whether this compensation is actually needed and what are the best values of the components.
– – – – – – – –
Mounting the sensor
First of all keep in mind that we haven't tested so maybe it won't work right. May not work well or not at all for a thousand reasons, could auto-oscillate or collect power frequency interference to the point of being completely useless. Unfortunately the little time we have is leaving everything to the NetHAL and the NetMaster that will work in network and even in WiFi (but not with all kinds of Master Pin and with a single Adc).
The disc that you see on the left is the face that goes toward the skin. The side towards the skin from now we will call “at the bottom”.
Of course the metallic parts must not touch the skin so it will take even a layer of thin plastic sheet.. A thin plastic like the one of the water bottles can fit. The plastic will “deep drawn” to hot to form a DRAM. Alternatively you might find a exact diameter short and very thin plastic.
The central part of the first disk is a sensitive area that needs to be coupled capacitively with the skin and which is terribly sensitive screen, and then goes as far as possible by electric fields at mains frequency coming from the electric system.
So around the sensitive area is called a ring “on call”.
The second disc is the posterior aspect of the first disk and need to screen electrically sensitive zone centrale from components of the third disc that otherwise would pass for ability their signals to record sensitive doing hopelessly autooscillare everything.
Around the first and second holes are metalized disk acting as a Faraday cage and help to shield the little signal that could have come from the sides.
Both the guard ring that the rear face are connected with a small wire which will pass through a hole to be made in silver rectangle “GUARD” the third disc. You connect them to “GUARD” instead of mass because otherwise the ability to ground completely would Dim the signal.
In the next image we see the silver rectangle where you will make the hole where it passes the wire that will be paid to “GUARD” the first disc. And also the rectangle “In” will have a small hole with a small wire that will be paid to “In” the first disc.
We didn't do the holes in plots GUARD and IN the third disk because you thought you mount it with the component side to the first disk and then do double sided with the back completely at the top and connected by copper by holes to its mass. But this Assembly would have made it impossible to do tests. For example,, change a component or solder a tad bit somewhere to make measurements.
Whereby the prototypes (yet to try) We did the third disc to a face alone and assemble with components to high #8217 &; (not to the skin). But be careful with this kind of installation you will have to pay attention to two things:
1) The outer ring of holes are not metallic and then do not touch the outer face of the first disk.
2) That the components will not be left uncovered but you must add a disk “all copper” to shield upwards. This last screen will be paid with a single side wire, so that you can uncover to access components and then close during the functional test.
And finally, at three pitches on the right: SIGNAL, + 5 and GND, settled three soft wire that will go to a Master Pin configured as Adc16.
– – – – – – – –
For those who want to experiment, Here's the project at present (as soon as we have time to improve)
Eagle project downloads + simulations + 3D images