pcompin

The electrodes were the biggest part to the construction of the EEG.  We had to create adn etch the boards ourselves from scratch.  Someone might stumble upon this blog at some point, so I will try to be detailed in describing the process and problems as well.

For starters we made Pedro Ortega’s active Electrodes, which were going to give us the strongest reading and we wouldnt have to get any electro-gels.  We liked this because A. it apparently gave us a better reading than the passive electrodes and B. it gave us the ability to just place the electrodes on the head instead of affixing them to the body with gels, etc.

We had to start collecting all the materials, which was a bit of a treasure hunt for us.  There was a lot of things we werent sure about how to get but after several days of hunting, we got the majority of it.
The electrodes needed:
a copper board, etching solution, container for etchant, 4 x 200k (we used 220K and it worked fine), 4 x 10k and 4 x 100Ohm resistors; 8 x 100pF and 4 x 10nF capacitors; 4 x TLC272 opamps; some silver wire of 1mm diameter

ETCHING:
So in terms of etching, there are a lot of options out there.   We tried to go “green” (literally) and get muriatic acid at the hardware store (~5$) and hydrogen peroxide.  It gets mixed in a 2 to 1 solution with the peroxide being the “2″ and the muriatic being the “1″. When you pour it in, it releases a gas (im assuming its CL gas) which can be a little toxic smelling, but dont worry, its only frying a little bit of your brain.

You take the layout for your PCB board (which for us, was conveniently on Pedros site) and print it out on MAGAZINE PAPER (FROM A PAGE, NOT THE COVER).  we got confused as it said “polyester paper” so we figured we could use any old fabric made paper (read: cotton) and it didnt work.  It needs to be water resistant so the LASER printer (which is also very important) ink sticks to it.  The laser printer toner is made from plastic and prevents the board from being etched where it is printed.  fairly cool actually.

clean it off with acetone, fyi, when its done.  its fast and and you don’t waste your time scrubbing.

DRILLING:
Ok, so you need a 1mm drill bit.  its a pain in the ass to find one and it needs to fit into a dremel.  We found one around the corner from NYU on 8th st at the hardware store next to the world of warcraft place. It takes a pretty steady hand to do it (which I don’t have at all) but luckily we didng break the bit.  You have to dripp about 36 holes per board so the process gets tedious.

SOLDERING:
the easy part, right? no.
The parts need to be SECURELY sodered to the board. then CHECK THEM with a multimeter.  This is important.  That way, when you go to cover them with silicone/glue them to your helmet/seal them with hot glue/whatever –  you know that they work on the board end of things.  Also, you will do a lot of soldering wires, so be careful not to overdo the soldering.  The copper will come off the board with time.

WIRING:
So this was the trickiest part. Our suggestion: go buy cheap XLR cables at a sound shop in sunset park and haggle with the guy on price.  two elctrodes will make one channel. So you ended up soldering an XLR to each electrode, then joining them together with a Y connector, then soldering that XLR to a 3 pin female connector. It gets confusing there, so be careful.  ****THIS IS IMPORTANT**** There are 4 connections on the electrode (-, +, an arrow, and a box).  The + should be your your power, the “-” should be your ground and should be connected to your SHIELD in your shileded XLR Cables.  The arrow should be connected to the other wire in the XLR.  the BOX is not connected to anything.

[View with PicLens]

We used an acrylic globe from Canal Plastics to make the “helmet” for the EEG.  We went for a weird sort of “steampunk”ish look for the device.  The electrodes looked so painful all the time (and they were) and we figured to carry that weirdness all the way through the design. We bought some plastic tubes and springs so we could give the electrodes pressure on the head so they would make full contact (and hurt really bad)

It was really important for us to make this look clean and interesting.  We had seen someone using baseball cubes in the workshop one day and thought it would work perfect for our project. So we got a couple of those, some sheet metal at home depot and got to work.  We needed to make a casing for the Olimex boards because the Analaog required we shield it in a metal box.  The digital could be in plastic, but they needed to be connected to eachother.  So we made a tiny metal box, cut some holes, and started drilling out the plastic softball holder.

[View with PicLens]

So the EEG project came and gone and I missed a bit of documentation on the blog. So I’m going to update as much as I can with the various sections of the project….

So I spent a lot of time thinking about what to do with this EEG machine.  I knew I wanted to create the sensors to catch these elusive brain waves, but what would I do with them when I have them? What would I actually be measuring in the first place? Would I actually get values that were polar enough and fluctuate based on thought process?

I had done a lot of research into what I was getting and the interpreation of the values. Basically, depending on WHERE I put the inputs on the head, I should get different types of values.  Things on the top should focus more on the motor cortex and thoughts like “physical movement” and “move body part” should potentially bump some values.. The actual moving of the arm will bump larger values.  Then there is a linguistic center that would spike values, light sensation, and just plain old emotional thought – happy, sad, frustration, calm, angry, confused.

Any other sort of communication beyond that is impossible. Even more so, the reality of this tool is based more in CONDITIONING and FEEDBACK of the brain than actually reading the “mind”.  My initial thought was to make an input device – like a computer mouse – that the device would control.  I like this idea, still.  However, the concept revolves around the user getting feedback from the device, practicing, and then ultimately controlling the device. Thus, it becomes an extension or a tool, but a much more complicated (granted, impressive) device than the original.  It has its base in assistive technology, which I love, but the practicality of it is a bit convoluted and perhaps out of my realm of expertise.

So I dug through some more ideas today..

Mouse Controller
Control two servos (hook it up to the midterm hand?!)
Control a scroll and a paintbrush
Control the color of a video
Fit the device into a hat and have emotional LED display
an media editor based on the emotions of the user.

The last one is the one that fascinates me the most.  the EEG should be able to recognize brain patterns instantly of a user in terms of calm, frantic, light and dark.  Even the basic toys out there do this.

Sometimes, sitting in all these programming and Pcomp classes, you forget what you were good at.  I focus on how I can’t figure out saving 2D arrays in proceessing or how the servo motor values need to be ironed out in order to work properly.  I forget that studied and worked for close to 10 years with media and video.  I forget that I still do this work and my goal in those projects is usually to CONTROL PEOPLES EMOTIONS.  Scary movies are made to scare people. Comedies are made to make people laugh.

So what I’m aiming for now is for the EEG reader to pickup base level emotions.  And either match or counteract the data with specific types of media aimed at controlling those levels.  If I’m in a dark room, could the machine recognize that and turn on a light, based on MY PERCEPTION, not the actual surroundings?  If I’m listening to bad music and the machine senses me getting more unclear or less focused, could the machine change the music for me, to something it knows puts me in a better place.

this sounds like that story we read 5 weeks ago. But I like it.  I think using the EEG as a literal “media controller” might be the next place I head with this.  Although, I’d like to build the EEG first to see what readings I am going to get from it, if any at all.  But I have to wait for all that to arrive. tuesday……..

Some sad news on the mouse front.

There is no easy way to hijack a USB mouse for use on a mac computer.  USBs have VID/PID (vendor and product IDs respectively) that they use to interface with the computer.  An older serial-PS/2 model mouse would work, but only for windows computer . I have Windows running on my mac, but it may just be all and all too complicated to work on while trying to simultaneously build a brain machine.  I took a part an optical mouse just to see what was up with it.. pictures below.  Things are pretty simplistic in it, but i have a feeling the microchip in this one is hiding under the sensor for the light.  I had read up and found that you can hijack mouse encoders to send the PS/2  or USB data to the computer, but these generally only work on windows computers anyway.

So thinking in another direction, I figured I could hijack a trackball mouse and just use servos and motors to control those physical inputs instead of attaching wires to them.  That would, oddly, be the most simplistic way to achieve a mouse, I’m guessing, at this point.

I also can just make a processing sketch to visualize a visual controller as a stand in for actually controlling the mouse.

So my options at this point are:

1. Use a PC (or virtualize) with Windows for a Serial/PS/2 connection in order to create a actual arduino to mouse interface.

2. hijack a trackball mouse with motors/servos and controll the physical WHEELS of the mouse that would otherwise be plugged into the computer, as normal.

3. create a visual interface in processing that could simulate the concept of mouse movement, without actual using the Operating System of the computer (other than to run processing).

4. abandon the mouse concept and go with something more simple (like a visulaizer or simplistic communication device).

Right now I am leaning towards 2 or 3. Number 2 would mean more time in constructing the mouse device, but could be cool to actually see it work.  Number 3 would be perfect for actual testing. So perhaps I will make 3 then, if time permits, option 2.

This is the current progress ive been making on things (click to see the larger verson)  Ive ordered the EEG parts($200!!!) and they are on there way.  They should arrive early next week.

Ive also been doing some research on hacking a mouse to work with the ardunio so I can prep for inputs and data from the EEG machine to push over to a “moused” arduino.. But more on that later.

I also have a column for design, which will come into play later as I get the EEG working, but im considering using another half dome acrylic sphere.  I liked putting it on my head when we made the robo-hand, so i think that was a pre-cursor to the EEG design.

I’ve decided to do my final project on inputs from a homemade EEG machine to create a Input device, like a computer mouse.

I’ve been doing a lot of research over the past couple weeks on how exactly the EEG should be made and possibilities that come from it.

First off, there are a lot of toys out there:

http://www.neurosky.com/

http://www.emotiv.com/epoc1.html

etc. etc.

Some are OK, at first, but I really want to get into the gritty of the EEG and see how the sensors work.  Plus, my research leads me to believe that the values coming out of them arent that great.  So I have ordered some parts from OLIMEX that are the basis for the openEEG plans for constructing an EEG (modularEEG).  The boards are an amplifier board and a digital board.

I talked with Neil from class and he joined in on the order of the boards, so we will build our EEGs together and probably end up collaborating on a project to some extent.  The EEGs are touchy and dangerous.  There are a lot of warnings out there about not touching your computer while wearing one because you might blow up your brain.  In the schematic, there is a ground saftey between the inputs for the head and the amplifiers with 5-12V of electricty. I’m not that worried, but this project could kill me. literally.

did the H bridge with the motor.. pretty easy stuff.. it looks like it spins pretty quickly too.  I’m wondering what I can find to attach to that and also how much it actually can pull. When we made the roboArm, Chris A. had discussed, after we had finished, using these motors instead of the Servos to wind up a rubber band instead of pulling a string.  I’m not quite sure the friction would have been strong enough to be as polar of an Off and On, but these things could definitely wind things up!