Message Boards Message Boards

Miscellaneous

Brain bio-sensor headsets

Toggle
Brain bio-sensor headsets
Answer
5/7/18 11:31 AM
A few years back there was a thread here about using EEG headsets to give meditators real-time or post-sitting feedback on their sits and perhaps some type of objective standard with which to measure their experience and/or progress. If I remember correctly the overall trend of the thread was that the consumer headsets weren't really reliable enough to be very useful and the software they came with wasn't too good either.

Does anybody know what the situation is like nowadays? Anybody having any fun/benefit/learning with these things?

I barely know anything about the neuroscience of meditation but on some basic level the idea of being able to get some type of objective data about one's...
  • "attention/concentration" level
  • "mindfulness" level
  • default network level?
  • positive affect (eg Metta / Karuna / Mudita) level?
...after or during a sit and then compare that to subsequent sits is appealing to me and seems that it could be very quickly useful for practice, provided the headsets can provide enough accuracy and so on. Somehow I doubt that attention or mindfulness correlate with a single objective brain-measurement so they might be hard to track without something more sophisticated but supposedly the software for these consumer headsets do track those two mental attributes.

Some questions:

Anybody know if those 4 stats are possible to measure with consumer headsets?
Could anybody point me to any relatively easily digestible info on research in this area that is not too technical or dense?
And does anybody know of any good/workable/pragmatic use cases for consumer headsets other than the suggestions I listed?

Any other experience/thoughts/interest people have with this stuff and want to share is very interesting, thanks a lot!

PS. This is also interesting: 3d printable open source modular headsets... http://openbci.com/

Two other headset manufacturers are: https://www.emotiv.com/ and https://store.neurosky.com/
Neurosky have a $99 headset called MindWaveMobile2 which seems to be the cheapest one available, anybody have any experience with it?

RE: Brain bio-sensor headsets
Answer
5/7/18 12:40 PM as a reply to Andrew S.
I got shortly obsessed with this awhile back.

My takeaways, iirc, fwiw: the consumer headsets that have built in algorithms with gross categories like "calm", "stressed", etc are silly and probably not worth it for people with intermediate to advanced skills in meditation. Maybe they can be motivating for beginners. Some of the heavier grade headsets like the OpenBCI (as you mentioned) that give you raw data might be interesting to play around with. I'm interested in trying this out in the future but I'm too broke to afford a proper setup at the moment. But, in the future, I imagine trying something like this: wear it for every daily sit, write a log of every sit. After some reasonable period of time (month?), go through the log and look for patterns. Try applying a reasonable set of custom tags to every sit like "restless", "energetic phenomena", "unusually concentrated", etc. Analyze the raw data (I would like to try machine learning on it). See if you can find correlations between your tags and your data. Etc.

Something like this,
https://www.gwern.net/Zeo

This would require 1) Having the time and money to afford the more powerful headset(s) 2) The technical skills to set it up, analyze the results, etc 3) The meditation skills to make it actually interesting 4) The interest in doing this.

I made this multireddit awhile back to monitor these topics conveniently,
https://www.reddit.com/user/dadakinda/m/quantifiedselfstuff/

Searching 'meditation' on there might be fruitful. Or, posting to some or all of them asking about meditation might be interesting. I doubt, though, you will find many people who meet the 4 criteria above.

If someone or some combination of people meet the 4 criteria above please do an experiment like this and post the results. I think it would be valuable. If you meet all the criteria but 2) I'm sure someone on the DhO or elsewhere will offer to try to help you with that (I will). Etc.

As far as existing neuroscience goes I'm not sure on the details. I know some correlations have been found. You might try here,
https://www.reddit.com/r/meditationpapers/top/?sort=top&t=all

Hope this is useful

RE: Brain bio-sensor headsets
Answer
5/7/18 12:46 PM as a reply to Andrew S.
I don't have any personal experience with it, but I found the anecdotal reports from TAGSync to be pretty convincing that they were headed towards something like neurofeedback with similar effects to insight practices.  There was also a subreddit collecting some more anecedotes. 

Daniel has suggested several times that the afterimage in the fire kasina provides near-realtime feedback on the quality of concentration, and I have found that to be true in my limited experience with it.

RE: Brain bio-sensor headsets
Answer
5/7/18 12:51 PM as a reply to JP.
Neat

Added to my multireddit. Gotta catch em all

RE: Brain bio-sensor headsets
Answer
5/8/18 4:27 AM as a reply to Dada Kind.
Dada Kind: Very useful! Thanks for all this info.

I have the time and interest but not yet the devices or knowledge, but I'll post on the DHO if I manage to get my hands on even a rudimentary headset and am able to play around with it. I don't think I'll be able to afford a good headset anytime soon but I might be tempted to just get the cheap neurosky.

I'm personally less interested right now in going all QuantifiedSelf on this and measuring data over time with tags etc (though would be more interested in that in the future), and more interested in a device that can give real time feedback and to look for patterns during a sit or post-sit and to see for correlations between subjective experiences and the data output for basic biofeedback reasons, but with a device that has enough granularity that, say, if my concentration is increasing over the course of a month that that would show up in the device output. I guess what I really want to do is just be able to watch my subjective experience transform in real time in a sitting and have that correlate properly in an output device that has enough granularity that as my skills deepen over weeks, the data continues to be relevant.

I'm kind of weary that something like the Neurosky mindwave mobile 2 ($99) with its prebuilt algorithms might not be able to do that, but this article makes it seem that the device should be able to measure enough data with enough granularity to be useful, as long as its accurate:
http://luciddreamingapp.com/neurosky-mindwave-mobile-headset-review-how-it-may-help-with-lucid-dreaming/

The article gives me some hope that at the very least, different mind states might correlate with different brainwave outputs and that could be interesting to experiment with at least.


JP Lewicke: Interesting re Fire Kasina, that makes sense, thanks for that tip!

RE: Brain bio-sensor headsets
Answer
5/8/18 9:24 AM as a reply to Andrew S.
I named the multireddit quantifiedselfstuff somewhat arbitrarily.

The reason, I believe, I concluded that trusting prebuilt algorithms isn't worth it is that 1) it's hard to verify the accuracy of the algorithm/readings 2) there is little incentive for the companies to actually make their algorithms/readings accurate.

An article comparing different EEG sets to test their accuracy would be ideal. I don't see that article at the moment. Although, here are some for sleep trackers:
http://sleep.cs.brown.edu/comparison/
https://medium.com/@matthewscd/quantifiedself-sleep-and-a-crappy-set-of-tools-e54d2b05f642

Doesn't look good to me. Clearly it would be hard to do a similar comparison for EEG headsets since you can only really wear 1 or 2 at a time. Although, I'm sure one could come up with something clever. Like, wear them in similar situations and compare aggregate stats like max amplitude, min amplitude, avg amplitude, blah blah. It's my guess most sets aren't very aggregate.

Which problem would you rather have:
1) Cheap, proprietary headset with questionable accuracy and opaque algorithms
2) Expensive, extensible headset with likely-accuracy but no convenience, usability, or interpretive help

1) has the potential to be completely useless outside of placebo. 2) at least always has the potential to be useful