EEG + CPAP

[ironhands]

How many would be interested in tracking EEG data along with their respiratory data?

It wouldn't take "much" effort to incorporate dry EEG sensors into the headgear.

I've been thinking of working on my own sleep tracking software using a commercial grade EEG unit, but, it'd run me about $2k to buy the hardware($300) and software dev kit ($1500), so if people wouldn't want it, I'm not going to break my bank just to follow my own sleep states

[ems]

Hi Iron... that's a lot of money for something that wouldn't be particularly relevant to me. Others may like the idea tho...

[Lazer1234]

I'm interested......

[ironhands]

It wouldn't cost that much for the end users, the bulk of the cost is the licensing for the software creation kit.

The end-user would only be paying like, $200 for the hardware and $10-20 for the software. I think it would be very useful to anyone also suffering from narcolepsy, CFS/fibro, or just generalized EDS that their apnea doesn't explain.

[JDS74]

I would be very interested.
Could you be more specific about the hardware required and the software development kit?

Are you proposing something akin to the now defunct ZEO?

[avi123]

Even if you get it could you interpret the data? Most Neurologists can't. Only those who specialized in "Sleep Medicine and Electroencephalography" could analyze the data for Sleep Disorders.

[ironhands]

Actually I've done a lot more research today, and the costs might actually be substantially lower to both the end-user, and the initial startup costs. I could likely get it down to <$100 with a 3 sensor package that could attach to standard CPAP headgear. Startup costs would likely be about the same, but, with no software dev kit, it would take a little longer to get going during development.

As for the tech, there's a few "chips" it'd need... an arduino or PIC with a USB connection, and the EEG chip, 3 dry electrode sensors that could hopefully be tucked into the existing CPAP headgear straps.

As far as interpreting the data, that's a different matter. It'd be able to record the data from the entire night, and then hopefully give a summary breakdown of the important parts, time spent in each phase, sleep onset, number of neurological arousals, etc.

In theory, might not actually be that difficult to pull off. If only I had the focus to do it, but I'm bloody tired lol

Yeah, I guess similar to ZEO, I hadn't seen that product before. Shame they're gone now. Were they any good for keeping tabs on narcolepsy?

[Julie]

Avi - I don't usually agree with you, but I sure do this time! He can't seriously think he'll be able to understand - to his benefit and not the opposite - what he sees.

[49er]

ironhands, I am definitely interested in what you are proposing. Sounds very interesting.

49er

[ironhands]

Avi - I don't usually agree with you, but I sure do this time! He can't seriously think he'll be able to understand - to his benefit and not the opposite - what he sees.

Julie.... There's nothing to understand, it would pass raw EEG data to a PC, not perform clinical assessments.

IF it were to distinguish between sleep states, and wake/sleep states, that is also relatively easy when you're only basing it from a single EEG channel, as the resolution wouldn't be high enough. It would be an approximation. When a device like this would be under $100, nobody should expect a medical grade diagnostic tool. The classifications for the stages of sleep are clinically standardized by EEG data. The math involved is freely available in dozens of text books.

Additionally, there is already software and freely available software routines to process and interpret this data, so it doesn't matter that you and Avi have no confidence in my ability to create a viable prototype.

[chunkyfrog]

I am watching the Melon EEG project, which far exceeded their cloud funding effort.
The interface appears to be simple and durable, unlike the Zeo, whose sensors needed periodic replacement.
Incorporating sensors into headgear is problematic, as too many different types of headgear exist.
Still, any product can get bogged down due to many business faults, and I'm not sure the Melon headband is ideal to sleep in.
I know the Zeo needed improvement, both with the interface design and with the dependence on online processing.
Many changes could have been made, beginning with the initial pricing, which doomed the zeo from the beginning.

[ironhands]

The current Emotiv insight looks decent. Still haven't picked up their old EPOC, which, in my understanding, is the most advanced one on the market, but completely useless for sleep.

The incorporation to existing headgear could be problematic, but the band/strap system most of the cheap ones are using now would likely be a better solution. No point starting a project if there's already one in the works, I'd be better off just writing an app for the Melon to handle sleep, if nobody is working on it currently. Their SDK would likely already be able to identify the stages, and does allow for raw EEG data. It's using the Neurosky chips, they seem to be the standard in budget EEG applications, and are really simple to work with. It's how I would have built the hardware as well. I should probably contact them about the SDK package before someone else starts an app like this.

[TheChuckster]

My final engineering project at school was pretty much a DIY EEG with real-time data. There's documentation online at:

http://people.ece.cornell.edu/land/cour ... m55_mj294/

[ironhands]

My final engineering project at school was pretty much a DIY EEG with real-time data. There's documentation online at:

http://people.ece.cornell.edu/land/cour ... m55_mj294/

Well done for the $75 budget!

I'd like to get a price on the basic Neurosky chip for the basic amplification processes. Between that and a typical arduino, as you demonstrated, getting the data to the PC for under $100 is simple. From there it's just a matter of determining sleep states. That's just "simple" number crunching. I wonder why the big names haven't added this sort of feature to their headgear; the data would be useful to many, and they could jack the price up a few hundred dollars for <$20 in raw materials. They wouldn't need anything more than the amplifier, since there's already a microprocessor in the CPAP anyway.

[chunkyfrog]

I love where this is going!