If there's one hill I'm determined to die on (given the information that I have anyway) is that any scientific finding extrapolating information from "brain waves" or "region activation" is little more than modern day phrenology. Of course this is not an original thought, though I find it alarming that it's the minority opinion seeing as EEG/fMRI augury suffers from blatant methodological and even statistical flaws.
There are too many sources to list but I think [0] has a good and varied collection of what I consider to be statistically informed thoughts on "cognition" research, see e.g. [1] [2] for a gentle introduction.
(I should like to point out that I recognize the usefulness of brain measurements in medical contexts; I'm just saying that using them to say anything about non-pathological perception or metaphysics has so far been fantastically unsuccessful.)
As someone who works in EEG, you may be surprised that I somewhat agree with you.
Let's start with where I am in agreement.
EEG or fMRI data only tells us about electrical activity in the brain, specifically the firing of neurons in different patterns and areas of the brain. We are yet to make the absolute connection of how that electrical activity translates into thought.
What I do find interesting is that in the age when we were learning about the movement of fluid and pumps, everything was taken from the understanding of how blood moved through the body and how we "worked" as a result of this fluid moving through our pipes.
As we learned more about electrical activity, we suddenly thought "oh! The pumps are moving necessary biological matter, but it's really the electrical impulses that are the important part. Even when we look at how the heart or any other muscle works, it's all about the electrical now.
So what is the next phase? As we understand more about quantum mechanics, will that be the next state of science that we attach to our understanding of how we work? Will it be something else?
However, the way I see this differently from you is that knowing how the pumps work, or knowing how the electrical systems work does are still valuable to us.
This isn't like the old days where we got sick from "humours".
So don't through the baby out with the bathwater. Understanding the electrical function of the brain is likely important in reaching our next level of understanding.
Now a somewhat unrelated, but tied to your thinking. I think this is what most people miss when they are talking about Nuralink and other implantable BCI. When people talk about "connecting our brains to the internet", we have to realize that we don't really have the ability to directly put information in. We can try to map electrical activity to requests, but I don't believe we have that in a high resolution method.
This doesn't mean BCI isn't valuable. If you have Parkinson's or other medical issue that can be treated with stimulation, it is absolutely valuable. But getting a monkey to play a game is still a long way from getting information in. Plus, a joystick is still easier, cheaper, and probably more fun than just staring at a screen without moving your body....assuming you have a body to move.
> EEG or fMRI data only tells us about electrical activity in the brain, specifically the firing of neurons in different patterns and areas of the brain. We are yet to make the absolute connection of how that electrical activity translates into thought.
To do that, I think we’d need much higher resolution imaging, both in space and in time.
It also says temporal resolution can be about a second. Whether that is sufficient isn’t clear to me. I think we know too little of the human brain to make a judgment on that. On the one hand, https://aiimpacts.org/rate-of-neuron-firing/ says a brain neuron, on average, fires less than 2 times a second. On the other hand https://www.frontiersin.org/articles/10.3389/fncel.2016.0023... talks of fast-spiking neurons that fire hundreds of times a second.
The current data is like trying to figure out how a CPU works from recordings that don’t show individual bits, but only show what parts get active when doing floating point calculations, when doing I/O, etc.
The 1970's BBC series, "The Body In Question" [0] explained the functioning of the body as understood through medical history and I found it very interesting that the latest technology was always used as a metaphor (and sometimes more) to expand understanding.
What you are experiencing is being present for the development of a major leap in technology. You'll see a lot of trial and error and near misses until someone cracks the big breakthrough of how we non-invasively read information in bulk from a mind.
Have a look at all the historical cynicism that surrounded things like voice tracks in motion pictures or consumer automobiles.
I don't deny that neuroscientists will discover something big in the future; I'm just saying that people should stop expecting that the next breakthrough will come from crude tools such as EEGs and fMRI, at least in conjunction with present-day computational power and statistical sophistication. These tools are useful enough without assigning an aura of metaphysical power to them; as things stand, billions of dollars are being wasted on adding noise, and what's worse is that the field has a track record of nefarious applications[0]. You could argue that the latter are not the scientists' fault, but if we keep promoting grandstanding pseudoscience for no clear reason, it's only a matter of time before somebody with both power and ill will is going to use it to harm people, all the while purporting to act in the name of "science."
I'm not sure I follow; by the same argument, measuring bumps on people's skulls can't be ruled out as not crucial to future science since it hasn't yielded a breakthrough yet.
The merits of a scientific methodology cannot possibly depend on the small chance that some hitherto unsuccessful approach might result in an unexpected breakthrough. Aside from the fact that anything might hold the key to the next big thing, such an approach would inevitably end up following whims and fashions more so than robust and sound scientific reasoning. All else being the same, priority should be given to research which is grounded in solid experiment design and a firm grasp on statistic.
> All else being the same, priority should be given to research which is grounded in solid experiment design and a firm grasp on statistic.
Agencies like DARPA exist primarily to counter this line of thinking. We built the atomic bomb by building a bunch of facilities out in the middle of nowhere and letting scientists just kind of make it up as they went along.
I hope you genuinely understand the difference between measuring bumps on ones head and well established scientific instruments that allow you to take measurements inside the body. MRIs are not some sort of witchcraft that should be distrusted.
> We built the atomic bomb by building a bunch of facilities out in the middle of nowhere and letting scientists just kind of make it up as they went along.
Physics in the years leading up to WWII had reached levels of accuracy and predictive power never before seen, so I think this example supports my argument if anything. Also the team there did not "just kind of make it up;" sure, they followed a relatively unconstrained creative process, but the foundations they were building upon were astoundingly firm.
> I hope you genuinely understand the difference between measuring bumps on ones head and well established scientific instruments that allow you to take measurements inside the body. MRIs are not some sort of witchcraft that should be distrusted.
I understand the difference, I was using skull bumps to prove my point by way of contradiction. I also understand the usefulness of MRIs. I'm just saying that they are useful within a certain domain, and cognitive science falls outside of that domain to a large extent (certainly to a much larger extent than sensationalistic pop-sci publications would have you believe).
It does seem reasonable that EEG can be used for a very, very crude type of mind reading. I believe it can show that you are in REM sleep or not in REM sleep, for example.
There are too many sources to list but I think [0] has a good and varied collection of what I consider to be statistically informed thoughts on "cognition" research, see e.g. [1] [2] for a gentle introduction.
(I should like to point out that I recognize the usefulness of brain measurements in medical contexts; I'm just saying that using them to say anything about non-pathological perception or metaphysics has so far been fantastically unsuccessful.)
[0] http://bactra.org/weblog/cat_cognition.html
[1] http://bactra.org/weblog/algae-2017-08.html#neuromania
[2] https://www.powells.com/book/neuromania-9780199591343