NeuroOn, the smart sleep mask: not a medical device, but better than a coin toss
alxd.org
alxd.org
A startup that I wish would happen would be a truly NON biased review organization - eg consumer reports but for the internet age.
Not just a consumer reports as digital first but an organization that is able to respond to the deluge of new products & product claims we see as consumers.
We were recently trying to buy a foam mattress and immediately got sucked into this huge vortex of misinformation to the point that I just gave up.
On the one side you have "mommy bloggers" yelling "tempurpedic offgasses give you CANCER!" and linking to a group of similar non-scientific studies.
On the other end is a group that says "latex is the only all natural foam!!!!" but then links to amazon products made in Thailand / etc full of mispellings / no oversight / etc.
Or the recent expose of how truffle oils (1) don't actually contain truffles.
It's insanity!
Clearly the FTC isn't the solution @ this point to oversee consumer protection.
Amazon is not a reviews service, nor is YouTube, yet this is where people turn. Nobody else seems to be taking the issue seriously, so we are.
The hardest part of building something like this is the chicken and egg problem, which is why I don't think there exists a good service for this yet. It's actually really hard to find a group of trustable product reviewers who stay engaged. People don't buy reviewable things that often, so you need another hook to get them to come back and sustain the ecosystem. This has been the hardest part of getting traction for our startup.
We're working hard on the problem. Hopefully you'll hear more about us soon.
There's a ton of bloggers and paid reviewers: http://www.productreviewmom.com/2013/11/how-to-become-produc.... All you need is a way to aggregate the one or two complaints / solid facts per review while filtering out all the fluff promotion verbiage. Mining truth, in other words.
It's probably lower-risk to target a specific segment and do in-depth reviews, e.g. https://examine.com/about/ is apparently quite profitable (http://www.solorwell.com/examine-com-is-grown-up-700000-in-t...), and Grokstyle is doing interior design (http://www.ldv.co/blog/2016/5/30/grokstyle-wins-ldv-vision-s... https://www.cs.cornell.edu/~sbell/pdf/siggraph2015-bell-bala...). But that sort of site will never cover everything, or even a significant fraction of products, so there's lots of room for value-added reviews. I envision a site like https://reviewmeta.com/, but with more emphasis on information extraction (& actual reviews) instead of machine learning (ML metrics don't help much when comparing products; to really corner the market the flow needs to be review site -> Amazon rather than the reverse).
They clearly spent some effort on this; it's all original content, nice pictures, good layout, etc. But as a person buying a garlic peeler, it's pretty much unusable due to the lack of editing; spelling errors such as "realy", the majority of the page taken up by a how-to, and tons of fluff such as "Some of the peelers presented more problems than they solved." There's about one paragraph per item.
Meanwhile, Amazon reviews are composed mostly of worse fluff ("amazing", "excellent", "I use it every day", etc.), and there's more of it. But there is also the occasional interesting complaint such as "The interior has to be cleaned after 2-3 uses due to garlic juice." and "The rolling motion gave my friend arthritis". Definitely more than a paragraph of info, maybe even two or three. So the info is there, it's just a problem of extracting it.
I have been already notified that I may have misinterpreted actigraphy publications I cited - I'll look into that tomorrow and gladly update the post with my findings :)
"it could try to wake the user, resulting in grogginess and complete lack of energy till the end of the day". you cite source [4], which is the wikipedia article on "sleep inertia" : https://en.wikipedia.org/wiki/Sleep_inertia
The wikipedia page doesn't say you'll have a "complete lack of energy till the end of the day", or anything close to that. it says "Morning sleep inertia may take several hours to dissipate" and "Sleep inertia is the impaired cognitive performance immediately upon awakening, which decays over tens of minutes", and "In the majority of cases, morning sleep inertia is experienced for 15 to 30 minutes after waking".
so my question is, whose idea was it to embellish using the phrases "complete lack of energy" and "for the rest of the day" ?
I've written a dozen versions of this article trying to accommodate different readers, including non-technical startup community. One of the main functions of NeuroOn is its alarm clock, which promises to "always wake you up in the best moment", which seems not to be true in ~31% of cases.
When I used direct wording about "sleep inertia", my test readers with no background in sleep science just said "Meh, it's better than without the mask anyway, isn't it?".
It's an honest mistake on my part, which only shows how hard science communication really is - and how hard is to choose a language in debunking bold claims.
I had huge hopes for NeuroOn - similar to what some people hoped for with Myo Armband - to finally have some customer grade EEG / EOG available for hacking and playing with.
When I realized that the company doesn't have a prototype working as they claimed - or at least won't show it to anybody and don't have any whitepapers, I got really irritated. There was nobody to properly debunk them. Journalists don't have the knowledge, researchers won't be able to publish that, most hackers won't be able to communicate the message properly.
So I decided to do an experiment myself, as open as possible, for everyone to recalculate and prove me wrong if I am.
You are probably looking for https://www.ncbi.nlm.nih.gov/pubmed/20374444
Lucid dreaming induction function in neuroon looks pretty neat though.
It's not standardized - so you'll have to make a headband yourself and choose some electrodes, but it gives you full access to all the signals. And don't worry about the 250/125Hz limitation, it can be quite easily hacked to write to SD with much better sampling rate.
I've played with OpenBCI and with some friends we tried to make it save signals with 1kHz sampling rate. No big-scale experiments so far.
Two days ago I checked out Myo on a friend born without a hand, with some hopes to create a simple interface based on her "phantom limb" gestures (she has full forearm with normal enough muscles). Sadly, Myo doesn't allow access to raw EMG signal - and my idea is to create Myo-like open hardware electrode boards to connect with OpenBCI and get some serious signals.
EDIT: Regarding OpenBCI - it's open hardware, you can make it yourself for ~under $100 if I'm right.
http://shop.openbci.com/collections/frontpage/products/myowa...
It's a little 3-lead EMG sensor intended for use with arduinos and the like. It's not quite as nice as the Myo (it uses sticky electrodes), but it's a similar form factor and much cheaper.
https://alxd.org/images/14_neuroon_signals/examination/elect...
What percentage of people have lucid dreaming using it? Are those people experienced lucid dreamer already? Just this aspect alone is enough to gain my interest.
Also how is it for side sleepers? Ever since my motorcycle accident I can only sleep on my side.
As someone who used NeuroOn I can't see anything wrong with using it sleeping on the side. It's still uncomfortable (to me), but it's not a deal-breaker.