For the last couple of months I have been working on a dream detector based on an EOG (electrooculogram). These devices are used to record eye position/movements in sleep studies etc. They work by picking up the voltage field generated in
the eye between the retina and cornea via electrodes. This voltage field changes with eye movement.
My current prototype provides audio feedback when the dream state is detected and measures about 30 by 50mm (1.2 by 2 inches), small enough to mount on the forehead while allowing
the wearer to sleep on their side. The sensing unit and electrodes are built into a headband.
Currently I am still working on the embedded software, but I have made a program which tests the pickup, sleeps for a set time then logs eye movements and gives feedback
when required.
I’ll post some pics, just need to borrow my mates digital camera.
The header is for programming the micro and downloading
the log, the headband has adjustable velcro mounting for the electrodes and fixing on the head. The little board at the
top of the third pic is the battery holder for the button cells.
As you can see the board is homemade (ie no fancy soldermasks etc) and all devices are surface mount.
The electrodes are designed to be used dry, no messy conductive
gel or skin prep. The pickup is working well at the moment, just need to do more work on the software side, especially on the setup menu’s etc which are accessed using the two buttons on the front. Either of these buttons will give a reality check beep when its running/sleeping. At the moment I have to adjust the settings by reprogramming it. As I said, more work required!
Looks nice I’ve never bothered with SMD components, seems to complicated even if one can make amazingly small devices
Doesn’t other face muscles interfere with the reading of they happen to move? This might happen during nonrem, then again I’m not sure… I built a novadreamer clone that used a simple photodiode and I found that it was more sensitive that I expected it to be, but that was easy to sort out in programming.
What microcontroller or cpu do you use? I used a pic in mine but since then I’ve “upgraded” to atmel. But now my atmel programmer seems to have give up even if it’s just wires and 3 resistors
Wow cool you must be really good with your technical engineering knowledge to make something so sophisticated, I hope it all goes really well for you. We need more research in this field, imagine haveing a pure remarkable device which could influence your mind
to recreate another world at any given time at the switch of a button. Its nice to see people trying and then hopefully succeeding
Yeah, movement does affect it, it’s quite sensitive. But from my logs it seems you don’t move that much during sleep, there are long periods of inactivity with a few isolated glitches here and there. Dreams themselves tend to build up and fade away in terms of REM intensity, and 20 REMS per minute are needed to trigger the feedback at the moment.
The micro is an ATMega8, and has ‘self programmable’ flash, allowing a log to be stored. I was actually thinking of using a PIC for the lower power consumption!
As for testing it’s not really up to it just yet, the user can’t change the settings. Working on it though, I’ll keep you guys posted.
hehe, I made mine so that I had to reprogram each time I wanted to change a setting
I made a sort of logger too but it only saved the time it sensed movement and in 5 minute intervalls since I had so little memory to work with But I did have it connected to a 16x1 lcd so I didn’t have to transfer the data to my computer to read it. I had the lcd on the wall and lots of cables going to the mask.
You made the electrodes yourself or? I’m actually getting interested in making something to log my remtimes. I found it quite hard to sleep with a big mask on the face so something like that would be much more comfortable to have on
And you have that headband so the electrodes are on the forehead or do they have to be on the side of your eyes?
The electrodes are made out of tin with a velcro backing for adjustable attachment to the headband. Tin is good because it doesn’t corrode, chlorided silver (or gold ! $$$!) would be better in terms of noise. Most commercial electrodes for EEG, EKG etc are one of these three materials.
As for results, well it works really well as a dream alarm, a la novadreamer. I’ve attempted to induce lucidity by giving low level feedback followed by the alarm a few minutes later. I’ve never noticed the cues in my dreams but the alarm is quite loud and wakes me every time it’s activated.
The log is very handy for confirming I’m dreaming when it’s activated - I find it hard to fully awaken and write my dreams down when the alarm goes off, pure laziness! I guess I have to adjust the initial feedback level to make it more effective as an LD induction device. Having said that I haven’t done any training to assist cue recognition.
The embedded software for the user setup is coming along, this is my third attempt at building a LD device and it works much better than the previous two. The first was based on a PIR (passive IR) pickup and the second the popular IR transmitter/receiver pair.
The main disadvantages I found with these is that they were very sensitive to shifting of the mask during sleep, as well as having to have sensors nearly above the eye, not great for comfort. Also, who wants an IR LED shining directly into their eye from a very close range? I believe these reasons were why the Novadreamer did not work that well and was withdrawn from the market, feel free to correct me. Rant over!
are you still working on this ?
it’s…so…small… amazing. THAT would be a good LD device… something fairly close to the skin that won’t bother you when you sleep face down
But I did have it connected to a 16x1 lcd so I didn’t have to transfer the data to my computer to read it. I had the lcd on the wall and lots of cables going to the mask.
amazing. THAT would be a good LD device… something fairly close to the skin that won’t bother you when you sleep face down 
