I love the approach from a science fiction angle starting w the classic tropes- frames the whole BCI discussion around the use cases and potential applications!
As for reading thoughts/feelings: There's 2 parts to that: reading someone's thoughts/feelings, and then feeding them to someone else. A lot depends on how you're picturing it.
For reading: feelings seems like in the realm of possibility -- most current BCIs are in the motor or sensory cortex, but if you stuck one in the amygdala (plus monitored hormone levels) you could probably pretty reliably read someone's emotions, and there might even be a fair amount of person-to-person similarity so you wouldn't have to completely retrain for each individual. But there definitely would be some learning involved, so there would be some long boring process of teaching the computer "I am feeling sad now". Gradually it could learn a possibly pretty nuanced mapping of your emotional states. Reading thoughts would be similar but a lot harder -- thoughts can be a lot more varied than emotions, there's higher dimensionality, but in principle you might be able to get a fuzzy picture of what someone's thinking.
Putting those emotions in someone else is very different. One version would just be to tell you "they're feeling sad now", without having you experience it. But if you want to actually induce the emotion in a brain, you're going to be very limited, because the patterns of activity associated with emotions are complex and our ability to induce activity are really blunt.
Ahh I see. So if the BCI reads, I'm sad based on a long process of learning what sad looks like in me, it would be able to communicate to your BCI (in theory) - Dom's sad.
Do you think it would be possible if your BCI learns what you're sad looks like, and mine learns what my sad looks like, that it would be able to see that I'm sad, and then induce the "sad" state within you? It sounds like the capability to induce a state is not currently great, but theoretically speaking that could happen eventually, or would you see further limitations?
The major obstacle is that our methods of inducing activity in the brain are really crude and just manipulate the broad activity in a region of the brain. We can't induce really specific small-scale patterns, which most complex brain states are. Even if we could induce very specific patterns (for example, with advances in optigenetics), the greater the complexity of what we're trying to induce, the more training you're going to need to characterize that pattern, and the greater the clarity of capturing the neural activity we're going to need.
For simple things like "general sadness", maybe there's a very simple neural representation that you could read and induce, but that "transfer" is going to be crude and miss all the nuances in the specific sadness you are feeling. Maybe in some far-off distant future where we have nanobots crawling all over every neuron we could do the super crazy stuff, but that's so speculative it's beyond what we have line of sight to. For any time in the medium-term future you're probably better off describing your emotion and relying on natural empathy ;)
Thanks for the shout out, Tommy! Great article, I really enjoyed it.
It's fascinating to ponder the possible future breakthroughs. I was recently chatting with a friend (who works in the area). She mentioned some of the new ways that optogenetics and robotics are being implemented. All I'm allowed to say is, watch this space.
I think anyone who is interested in this will benefit from reading up on Transhumanism. It talks about the benefits and ethics of enhancing the human condition but also how it could go wrong very quickly.
Currently, the biggest use for invasive BCIs are medical -- people who have paralysis, ALS, or other disabilities. If it meant being able to use a computer when I otherwise would be confined to bed with nothing to do, I would certainly consider a brain implant!
So would I. But there's nothing putting the brakes on this - where and when does it end? As I said, the brain is an enigma. We don't know the long term effect of implants, and it's irresponsible to use them to alleviate some people's boredom.
Why should it end? If there exist people willing to try, willing to take a chance while fully aware of the risks involved, then why should they be stopped?
The brain will only remain an enigma if we stop studying it. At a certain point, that means experimenting with the sorts of technologies.
Very interesting article. I can't help but feel I will be one of those Luddites who abstains from this kind of brain/computer interfacing. Like the friend of the protagonist in the Black Mirror episode 'The Entire History Of You', who at a dinner party reveals she has abstained from getting the implant. Perplexed, everyone else speaks to her like she is Amish or something, unable to understand how anyone would want to live 'naturally'. I always resonated with that character the most.
Nice discussion of Brain Computer Interfaces (BCIs), @Tommy Blanchard.
It is only natural, should human technology progress, that we need to integrate that technology directly into our brains.
Not only could this help shield us from the possibility of being “left behind” by AI, it would greatly increase the bandwidth between us and the machines we use everyday.
I don’t know about you, but while I cannot live without the supercomputer in my pocket, I find the touch screen incredible slow and clunky.
As I understand it a person with the most radical BCI setup today needn’t learn anything in order for the technology to work. Instead it’s sometimes the computer that must learn how that person’s brain works. I’m referring of course to recent experiments where a person with atrophied speech muscles would try to speak certain known text for 100 training hours, and a computer would be set up to potentially determine what she was trying to say by crunching this against detected electromagnetic radiation from certain specific parts of her cerebral cortex. Apparently it worked pretty well, so things can also go that way too — the computer learns you rather than you try to interpret something different from what’s standard.
Secondly I wonder if you’d comment on what I think this suggests about consciousness itself? Would we expect any aspect of the electromagnetic field associated with a standard computer to correlate very well with what its screen’s pixels are suppose to be doing? Unless specifically designed to do so (as in the case of bluetooth for example), I don’t think so. Standard computer EFM seems to be functional garbage that the computer must be protected from in order to function right. So why was it possible for brain EMF to be transcribed into the words that she was trying to say? It seems to me that this should be because her (and by extension our), consciousness probably exist electromagnetically. Any thoughts on that possibility?
Good point that for some applications (like language) the learning will be on the device's side. But there's still a long, active learning process before it can be used. If it's the computer learning, the person is still involved giving feedback on what they are thinking (or more likely, being told what to think)
In terms of consciousness, I don't think there's much that this tells us about it. We know the brain processes information with electrical signals. We can read off some of that information by detecting some of that electrical signaling. No need to jump to specific theories of consciousness to account for that
Ah, well I also have many other reasons to suspect that consciousness exists as a neuron produced electromagnetic field! I don’t know if you have any familiarity with Johnjoe McFadden’s EMF consciousness proposal, but I suspect that empirical evidence will make it uncontroversially accepted in the mainstream some day. The problem being that today the field is riddled with all sorts of unfalsifiable nonsense and thus testable theories like his are virtually unheard of. Before these new BCI experiments, if he were more on the ball I think he’d have seriously raised some eyebrows by predicting that it ought to be possible for detected EMF at very specific locations of the cerebral cortex, to be correlated with the words that someone is trying to speak, and specifically because he proposes that such EMF causes speech muscles to operate as they do my means of associated ephaptic coupling. But alas, he said nothing.
Instead of just EMF detection, I’d like researchers to add EMF transmitters to someone’s brain that are set up to produce energies that mimic the energies produced by standard synchronous neuron firing. If such energies exist as all that we see, feel, smell, think, and so on, then we’d expect an experiment’s exogenous energies to constructively and destructively interact with that consciousness. Thus a subject ought to tell us that their consciousness gets whacky in specific ways with specific types of energies, and not knowing exactly when or which energies are being produced. But if researchers are able to produce all manners of appropriate energies centered at various locations of the brain, though the person doesn’t notice anything strange, then consciousness probably doesn’t exist as a neurally produced electromagnetic field. In this sorry field, in that case wouldn’t it at least be wonderful to finally have one demonstration which displays how science is suppose to work?
Thanks for the shout out @tommy!
I love the approach from a science fiction angle starting w the classic tropes- frames the whole BCI discussion around the use cases and potential applications!
Nice piece - thank you for this!
What do you make of the sci-fi capability to read each other's thoughts, or feel each other's feelings with BCIs?
Thanks!
As for reading thoughts/feelings: There's 2 parts to that: reading someone's thoughts/feelings, and then feeding them to someone else. A lot depends on how you're picturing it.
For reading: feelings seems like in the realm of possibility -- most current BCIs are in the motor or sensory cortex, but if you stuck one in the amygdala (plus monitored hormone levels) you could probably pretty reliably read someone's emotions, and there might even be a fair amount of person-to-person similarity so you wouldn't have to completely retrain for each individual. But there definitely would be some learning involved, so there would be some long boring process of teaching the computer "I am feeling sad now". Gradually it could learn a possibly pretty nuanced mapping of your emotional states. Reading thoughts would be similar but a lot harder -- thoughts can be a lot more varied than emotions, there's higher dimensionality, but in principle you might be able to get a fuzzy picture of what someone's thinking.
Putting those emotions in someone else is very different. One version would just be to tell you "they're feeling sad now", without having you experience it. But if you want to actually induce the emotion in a brain, you're going to be very limited, because the patterns of activity associated with emotions are complex and our ability to induce activity are really blunt.
Ahh I see. So if the BCI reads, I'm sad based on a long process of learning what sad looks like in me, it would be able to communicate to your BCI (in theory) - Dom's sad.
Do you think it would be possible if your BCI learns what you're sad looks like, and mine learns what my sad looks like, that it would be able to see that I'm sad, and then induce the "sad" state within you? It sounds like the capability to induce a state is not currently great, but theoretically speaking that could happen eventually, or would you see further limitations?
The major obstacle is that our methods of inducing activity in the brain are really crude and just manipulate the broad activity in a region of the brain. We can't induce really specific small-scale patterns, which most complex brain states are. Even if we could induce very specific patterns (for example, with advances in optigenetics), the greater the complexity of what we're trying to induce, the more training you're going to need to characterize that pattern, and the greater the clarity of capturing the neural activity we're going to need.
For simple things like "general sadness", maybe there's a very simple neural representation that you could read and induce, but that "transfer" is going to be crude and miss all the nuances in the specific sadness you are feeling. Maybe in some far-off distant future where we have nanobots crawling all over every neuron we could do the super crazy stuff, but that's so speculative it's beyond what we have line of sight to. For any time in the medium-term future you're probably better off describing your emotion and relying on natural empathy ;)
Fascinating article, thanks!
Thanks for the shout out, Tommy! Great article, I really enjoyed it.
It's fascinating to ponder the possible future breakthroughs. I was recently chatting with a friend (who works in the area). She mentioned some of the new ways that optogenetics and robotics are being implemented. All I'm allowed to say is, watch this space.
👀
I think anyone who is interested in this will benefit from reading up on Transhumanism. It talks about the benefits and ethics of enhancing the human condition but also how it could go wrong very quickly.
Super interesting read, I enjoyed it immensely.
No. Just no.
😂
But since we actually know so little about the human brain, why would anyone want their brain messed with?
Currently, the biggest use for invasive BCIs are medical -- people who have paralysis, ALS, or other disabilities. If it meant being able to use a computer when I otherwise would be confined to bed with nothing to do, I would certainly consider a brain implant!
So would I. But there's nothing putting the brakes on this - where and when does it end? As I said, the brain is an enigma. We don't know the long term effect of implants, and it's irresponsible to use them to alleviate some people's boredom.
Why should it end? If there exist people willing to try, willing to take a chance while fully aware of the risks involved, then why should they be stopped?
The brain will only remain an enigma if we stop studying it. At a certain point, that means experimenting with the sorts of technologies.
Good job, man. Great article. You really nailed the psychology behind “human upgrading.”
Thank you!
Welcome! :-)
Very interesting article. I can't help but feel I will be one of those Luddites who abstains from this kind of brain/computer interfacing. Like the friend of the protagonist in the Black Mirror episode 'The Entire History Of You', who at a dinner party reveals she has abstained from getting the implant. Perplexed, everyone else speaks to her like she is Amish or something, unable to understand how anyone would want to live 'naturally'. I always resonated with that character the most.
Nice discussion of Brain Computer Interfaces (BCIs), @Tommy Blanchard.
It is only natural, should human technology progress, that we need to integrate that technology directly into our brains.
Not only could this help shield us from the possibility of being “left behind” by AI, it would greatly increase the bandwidth between us and the machines we use everyday.
I don’t know about you, but while I cannot live without the supercomputer in my pocket, I find the touch screen incredible slow and clunky.
Nice to meet you Tommy. A couple of things.
As I understand it a person with the most radical BCI setup today needn’t learn anything in order for the technology to work. Instead it’s sometimes the computer that must learn how that person’s brain works. I’m referring of course to recent experiments where a person with atrophied speech muscles would try to speak certain known text for 100 training hours, and a computer would be set up to potentially determine what she was trying to say by crunching this against detected electromagnetic radiation from certain specific parts of her cerebral cortex. Apparently it worked pretty well, so things can also go that way too — the computer learns you rather than you try to interpret something different from what’s standard.
Secondly I wonder if you’d comment on what I think this suggests about consciousness itself? Would we expect any aspect of the electromagnetic field associated with a standard computer to correlate very well with what its screen’s pixels are suppose to be doing? Unless specifically designed to do so (as in the case of bluetooth for example), I don’t think so. Standard computer EFM seems to be functional garbage that the computer must be protected from in order to function right. So why was it possible for brain EMF to be transcribed into the words that she was trying to say? It seems to me that this should be because her (and by extension our), consciousness probably exist electromagnetically. Any thoughts on that possibility?
Good point that for some applications (like language) the learning will be on the device's side. But there's still a long, active learning process before it can be used. If it's the computer learning, the person is still involved giving feedback on what they are thinking (or more likely, being told what to think)
In terms of consciousness, I don't think there's much that this tells us about it. We know the brain processes information with electrical signals. We can read off some of that information by detecting some of that electrical signaling. No need to jump to specific theories of consciousness to account for that
Ah, well I also have many other reasons to suspect that consciousness exists as a neuron produced electromagnetic field! I don’t know if you have any familiarity with Johnjoe McFadden’s EMF consciousness proposal, but I suspect that empirical evidence will make it uncontroversially accepted in the mainstream some day. The problem being that today the field is riddled with all sorts of unfalsifiable nonsense and thus testable theories like his are virtually unheard of. Before these new BCI experiments, if he were more on the ball I think he’d have seriously raised some eyebrows by predicting that it ought to be possible for detected EMF at very specific locations of the cerebral cortex, to be correlated with the words that someone is trying to speak, and specifically because he proposes that such EMF causes speech muscles to operate as they do my means of associated ephaptic coupling. But alas, he said nothing.
Instead of just EMF detection, I’d like researchers to add EMF transmitters to someone’s brain that are set up to produce energies that mimic the energies produced by standard synchronous neuron firing. If such energies exist as all that we see, feel, smell, think, and so on, then we’d expect an experiment’s exogenous energies to constructively and destructively interact with that consciousness. Thus a subject ought to tell us that their consciousness gets whacky in specific ways with specific types of energies, and not knowing exactly when or which energies are being produced. But if researchers are able to produce all manners of appropriate energies centered at various locations of the brain, though the person doesn’t notice anything strange, then consciousness probably doesn’t exist as a neurally produced electromagnetic field. In this sorry field, in that case wouldn’t it at least be wonderful to finally have one demonstration which displays how science is suppose to work?