Brain Monitors Are Going Mainstream, Despite Skepticism | Hacker News Reader

Brain Monitors Are Going Mainstream, Despite Skepticism

wired.com

20 pointsprostoalex11 comments

EEG is a fundamentally noisy and low bandwidth measurement of neural activity. I work with neural recordings based on intracortical electrodes with resolution of single neurons. I think of EEG as analogous to listening to an orchestra from outside the building. You may hear that there's a bit more brass than strings, and you might get some sense of the overall sound level, but you won't really be hearing the music. To strain this analogy, any time you move your facial muscles, it's like a truck driving by swamping everything out.

I work in a lab with EEG and FMRI primarily writing software for neural feedback which involves a lot dsp and physics, and considering that spike timing is a predominant theory in (comp)neuroscience today (and has been for nearly 40 years, yet no more fundamental understanding… go figure…), I think it would be an understatement to say that the level mathematical and physics rigor/theory applied by most people in the field is considerably lacking… if not a joke compared to what physicists/astrophysicists have had to deal with and have proven by experimentation in much nosier environments in the same time frame. I really see the field bifurcating with people doing more of the same, and those who want to import experience of modern physics…

I am aware of theoretical physicists applying statistical mechanical principles to understand neurons and biological neural networks (one example: [1]). This is a reasonably hot topic in bio-physics. I am also aware of researchers who are applying state of the art high-dimensional data analysis to MRI, etc. (one example: [2]).

I would modify your assertion to: "this kind of application of high-level physics and math should be more common."

[1] https://web.stanford.edu/dept/app-physics/cgi-bin/person/sur...

[2] http://jelena.ece.cmu.edu/bimagic.html

That is certainly a more diplomatic way of putting it, and I will try to use such language in the future because its probably a lot more conductive in exploring such with those who are interested but reluctant.

These really gross, noisy EEGs are useful for some things, but they're being peddled alongside all manner of snake oil. The structure of thought and emotion is so complex and obscure that leading scientists are barely able to even define it, much less detect it with powerful, invasive tech — much less again with this type of device.

If you find using an EEG device improves your life in some way, that's great - go for it. But exercise skepticism when it comes to the claims being made about just what is being detected here and with what level of accuracy and precision.

What kind of data can monitors like this actually collect out-of-the-box? I know that you can train a system to recognize some pretty high-level abstractions based on individual neural activity, but how reliably would a similar pattern-recognition mechanism work across a larger population?

These consumer-targeted devices are predominantly used to detect and act on changes in the relative power of certain frequency bands from rhythmic neural activity. These changes in activity are common across people, and don't require any specialized user-specific techniques to identify. For instance, alpha wave (8-12 Hz) power is correlated with arousal/attentiveness.

Interesting. So am I right to assume that these devices can detect a range of emotions? It would be fascinating to incorporate some kind of iterative feedback loop into a work of art.

No, emotions are extremely complex phenomena and can't be detected with this type of hardware.

Exactly, you'd need a mood ring for that.

This seems neat, but I'd like to see a study some sort on what sort of benefits (if any) it has over the long term.

It would be interesting to see if developers of these kinds of devices can eventually progress to one reliably driving a keyboard at anything approaching normal typing speed.

Reply on news.ycombinator.com