CLEARER THINKING

with Spencer Greenberg
the podcast about ideas that matter

Episode 075: Major and minor scales of consciousness (with Andrés Gomez Emilsson)

October 14, 2021

Should pleasure and pain be measured on logarithmic scales? How might such a scale affect utilitarian calculations? How do harmonic energy waves in the brain correspond to states of (or intensities of) consciousness? What sorts of conclusions can we draw about brain states given the resolutions and sampling rates of tools like fMRI, EEG, and MEG? What is the symmetry theory of homeostatic regulation, and how does it connect to pleasure and pain? Are uncomfortable or confused mental states computationally useful to us? To what extent can the concepts of musical consonance and dissonance map onto energy states in the brain?

Andrés Gomez Emilsson has a Master's Degree in Psychology with an emphasis in computational models from Stanford and a professional background in graph theory, statistics, and affective science. Andrés was also the co-founder of the Stanford Transhumanist Association and first place winner of the Norway Math Olympiad. His work at QRI ranges from algorithm design, to psychedelic theory, to neurotechnology development, to mapping and studying the computational properties of consciousness. Andrés blogs at qualiacomputing.com.

JOSH: Hello, and welcome to Clearer Thinking with Spencer Greenberg, the podcast about ideas that matter. I'm Josh Castle, the producer of the podcast and I'm so glad you've joined us today. In this episode, Spencer speaks with Andres Gomez Emilsson, about valence, harmonics, symmetry, and the neural and nervous systems that can pleasure pain and other states of consciousness.

SPENCER: Andres, welcome.

ANDRES: Thank you. Thank you so much. Happy to be here.

SPENCER: Yeah, happy to have you here. You're one of the few people that I've referred to as a modern mad scientist, I don't know whether you accept that naming or you reject it?

ANDRES: Not only do I accept it, but I also embrace it fully.

SPENCER: Okay, I'm sure this is gonna be a very fun and free conversation. So I want to start talking about the logarithmic scales of pleasure and pain. So what is that? Why do you think that pleasure and pain are logarithmic? And why should we care?

ANDRES: Yeah, yeah, this is a fascinating topic. And it's something that I spent quite a bit of time thinking about since 2019. We conducted a study and did a lot of literature review for this purpose, and ended up writing an article basically called Logarithmic Scales of Pleasure and Pain. We posted it into the EA forum, it's also on the QALY Research Institute website. And, in brief, all of these, like started out with kind of learning about super intense states of consciousness and kind of like reckoning that ethics textbook, or even EA forum posts, or even LessWrong posts, we don't really see the kind of like mentioned or taken into account as kind of a part of a strategy of prioritization, like, that seems to be very much missing. And then if you look at, for example, quality-adjusted life years as a paradigm, you don't really get to have extremely negative values, like, all you get is this life worth living, as a maximum of 100% versus 0%. But there's very little consideration about whether if this person is in a very bad state of consciousness, and this other person isn't even versus the state of consciousness, is there may be a very substantial difference, depending on just how actually bad they are. And conversely, there are different practices, different approaches to life, that may result in states of consciousness with orders of magnitude of different feelings of meaningfulness and sense of love. Like maybe that also matters quite a bit. And we have bumped into the people who can act as the Buddhists Janus, which are states of very high concentration. We also had talked to people who had significant experiences, not only with psychedelics, which obviously fit this description of very, very high-intensity consciousness but especially like 5-Meo-DiPT, basically, the reports that we were getting from them and interviewing them were quite extraordinary. And finally, also, I've talked to people who have epilepsy and talked about their very exotic, extreme states of consciousness. And I figured that this is not yet EA cannon, like, this is not something that is on the map as something that we should be considered for prioritization. So basically we decided to spend several months just focused on investigating whether we can build the case that actually pleasure and pain follow the long-tail distribution. And I started out very open-minded about what is the true scale and after doing a lot of the research and analyzing surveys, I give the other side a pretty convinced that logarithmic scales is the right lens to look at the valence. And I'd be happy to go into the reasoning and the background knowledge for this but that's at a high level, that's what it is.

SPENCER: So I think what you're saying is that when we think about people suffering or happiness, we tend to assume that nobody is a thousand times happier than the other person, nobody is suffering 10,000 times more than another person. And if we did believe that, if we believe that there were these kinds of enormous outlier states, it will change how we think about trying to minimize total suffering, we might want to say emphasize reducing these absolutely, incredibly bad states, rather than trying to make everyone suddenly less miserable or something like that. Is that correct?

ANDRES: Yeah, that's correct. And, I mean, if you think about it, a lot of utilitarian reasoning if you buy into aggregation, utilitarianism where you can aggregate values across people, and so on. Typically, they analyze a case where they minimize the average pain of people. But the thing is that if the scale is actually logarithmic or let's say, quadratic, just like not linear, just trying to minimize the average is not actually going to minimize the thing the scale is measuring, right? You need to basically correct for whatever non-linearity is actually in the scale.

SPENCER: Right. And when you say logarithmic, are you referring to a log-normal distribution?

ANDRES: Yeah, so there are two things here. One is the shape of the distribution of the intensity of valence of moments of experience. And I would argue that follows a log-normal distribution. And then there's how do you represent that.

SPENCER: Let's just explain log intervals for this that we don't know. Imagine that it's a normal distribution, except that it has a much fatter tail on one side. So it's much more prone to huge outliers. And then usually, the other side of it is kind of clipped off. So it can't go below a certain value. And I think you're saying that you have one log-normal distribution for pleasure states and a different log-normal distribution for pain states, and then you can have both simultaneously. So you could do both a pleasure and pain, say at the same time, is that right?

ANDRES: Well, would it be in a sense to arrive at is that you have a log-normal distribution for pleasure, and you have a log-normal distribution for pain. And then you have an interesting mixture of distribution that we don't really understand at the moment because we haven't studied which would be of mixed states. There are also states of consciousness where you have very intense pleasure mixed in with very intense pain. But for the most part, the things that we were examining were basically are the pure positive extremes and the pure negative extremes, and what kind of scale would best describe those.

SPENCER: Got it. Okay, cool. And then you were about to explain, there's another kind of when you say logarithmic, there's something else you're referring to. You want to go into that?

ANDRES: Yeah. Which is basically when you ask a person from 0 to 10, how much pain are you experiencing? But obviously, people's responses to this question will vary depending on their previous experiences. Their "pain tolerance", like their emotional state, and so on. So there are a lot of things that contribute to a response in that. But we found that, in a sense, for a given person, it seems that the general pattern of behavior is that constant increments, in that scale, map on to basic differences in proportion, as opposed to actually kind of differences in absolute values of the intensity.

SPENCER: So going from like a seven to an eight, might be twice as bad. Is that idea?

ANDRES: Yeah, yeah, exactly. I mean, this can be like an intuitive case, very simply by asking questions, like, you're going to be in two different days: in one day, you're going to be in 9 out of 10 pain, and another day, you're going to be in 1 out of 10 happiness or something like that, would you rather takedown from 9 out of 10 pain to 8 out of 10 pain? Or would you rather increase the day from like, 1 pleasure to 2 out of 10 pleasure? People will generally tend to care a lot more about the differences in the extreme of the scale. And this is intuitive, but then, if you actually ask people to have a great experience that they have had in the past, and then you ask them to basically describe the difference in intensity, like what proportion of intensity, what difference was, you will see that the kind of distribution that best fits, that the kind of data is something much more like a log-normal distribution. And that, generally speaking, rather than a fixed amount, say there's a difference of two points in the pain scale, it was like always the same kind of like difference in intensity. That is the thing that we think it's not the case.

SPENCER: Right. It's what do you think some of the strongest arguments are? That pain is logarithmic.

ANDRES: For me, there are a lot of converging lines of evidence, even from first-person experience, things that I have experienced that pretty much convinced me of, at least, a powerful kind of long tail of intensity of experience, at the very least. But I'll just mention, let's say three key examples. So are pieces of evidence. So the first one is, this is from very dry neuroscience. Basically, if you examine patterns of neural activity, you're measuring the activity of neurons, either in a petri dish or in vivo in some animal, and you're monitoring the activity of lots of neurons that are spatially close together. And what you will find is that it's not that each of these neurons is firing, kind of like randomly, maybe with a personal distribution or a Gaussian distribution. It's not like these neurons are just following a normal distribution. But rather, they seem to follow these patterns of cascading activity, something that in neuroscience is described as exhibiting a criticality, meaning that small events can basically cascade or avalanche into very large events. And that if you plot events of activity, you will find that it essentially follows a log-normal distribution. And this is very, very similar to, if you actually were to monitor avalanches in various countries or in various mountains. You will see that most avalanches are very small, they're kind of just like little updates and adjustments in the distribution of snow in the mountain. And every once in a while you get this insane, incredibly massive avalanches; something you can do, for example, like Google the most deadly avalanches and you will see that the most deadly avalanche of all time, like killed something like 22,000 people, just far, far, far more deadly than the runner up, and so on. There are going to be these very big steps. An analogy that I would make here is that this is a function of the weather, when you have in a sense, the right set of conditions for a lot of snow, you will get like a ton of avalanches, and you need all those conditions to happen. And if you don't have them, then you may have just you know, zero, or maybe one or very few, for the same reason, in a sense. If you plot the number of avalanches in each country, you also get a log-normal distribution. And likewise, I just have many ways of slicing and dicing the data about the size of avalanches and where they appear, how long do they last, and all of them tend to show a log-normal distribution. The same, I would say, applies to nervous systems, some nervous systems, if they're in a highly irritated state, or maybe a state after enduring chronic pain or something like that, i's in this sensitized state, and you can think of it as the neuronal environment of that state. And in a sense, that will be a very avalanche-prone state. So you get these very large cascades of activity, which would be kind of these events are like spikes of pain. In this particular case, the big picture is that you know, even when you're examining a moment, or like an event where somebody is experiencing intensive suffering, even within that, that moment, you know, within the span of a few seconds, you will see that there's kind of like these huge spikes that happen like that people rarely report something like the pain was perfectly constant. Like it's almost kind of like inherently, in the quality of pain, there is kind of this roughness and irregularities that come with it. And in the sense that yeah, there's I think, like the analogy with a kind of avalanches and other systems that exhibit criticality can be pretty solid, I guess. Yeah, that's, that's one piece of evidence. And it's ultimately grounded on dry neuroscience analysis of the distribution of neural activity as a function of the neuronal environment and as a function of the stimulation patterns that you give to it.

SPENCER: Okay, so we have this very fat tail distribution of neural activity. And maybe we could say that that maps on to the kind of internal experience you have, what do you think another strong line of argument is in favor of this idea of logarithmic pain and pleasure scales?

ANDRES: Yeah, the other line of evidence is when you look at basic people who have examined very, very negative states of consciousness, in one way or another. Most of them end up actually creating scales that are explicitly logarithmic. So for example, there's this thing called the KIP scale, which is to measures the intensity of cluster headaches, which are basically one of the absolutely most painful things that people can endure. And it goes from 1 to 10. And the author is like, very explicit that in a sense, like the difference between a 5 in the KIP scale and a 10 in KIP scale is a difference in an X, like, it's a cluster headache is 10 times more intense, more bright, more energetic, when you're having a 10 out of 10 rather than 5 out of 10. And talking to people who suffer from cluster headaches, they all agree that like yeah, this is the appropriate way of describing cluster headache intensity. Importantly, you also see this affecting, for example, what's called the Schmidt Pain Index. So there's Justin O. Schmidt, a really interesting entomologist who decided to basically catalog the kind of stings that different insects have a particular class of insects, basically the class of wasps.

SPENCER: He did this by getting stung a whole bunch, right?

ANDRES: Yeah, actually, yeah, he would endure it himself a bunch of things. And basically, he standardized kind of the scale. So there would be very crazy things in some situations, he would get stung with 10 bees, for example. So that he had in the repertoire of his experiences. Okay, what is it like to have 10 stings from bees, you know, cluster together in your arm, and then compare that with a single thing from a harvester ant. And then by comparing is like, Oh, actually, the harvester ant is equivalent to the intensity of 10 stings from bees. Basically, he has like this four-point scale. And in his scale explicitly, he basically says that each increment in the scale is a multiplicative factor by 10. So the difference between a 1 and a 4 in the Schmidt Pain Index is a difference of 1000 in the intensity of the pain.

SPENCER: I think this scale is really, really amazing. I just want to read some examples because he actually describes all the different stings which I learned about from your article. So I'll give you some of his descriptions starting at a 1 then at 1.5 and so on. So at the one pain scale, we have the Indian jumping ant which he describes as "Ah, that wonderful wake-up feeling like coffee but also bitter." Then for 1.5, we have the ferocious Polybia wasp, which he describes as, "like a trick gone wrong, your posture is a target for a BB gun, a bull's eye over and over." Then we have a 2, the Western honeybee, which he describes as "burning corrosive but you can handle it, a flaming match had landed on your arm and is quenched first with lye then was sulfuric acid." Okay, now we have a 2.5, the yellow fire wasp, in "a distressing pain, tiny blow torches, kiss your arms and legs." Okay, then we have a 3, getting almost the top of scale here, the red paper wasp, "caustic and burning, distinctly bitter aftertaste, like spilling a beaker of hydrochloric acid on a paper cut". And then finally we have a 4, the top of the scale, we have the bullet ant, "pure intense brilliant pain, like walking over flaming charcoal with a three-inch nail embedded in your heel". So that gives you some idea of how he describes these things.

ANDRES: Yeah, it's kind of terrifying. And the thing is it really doesn't end there. I mean it does seem like there are actually a couple of things that are worse. There's a particular jellyfish that produces something that is even worse than the bullet ants. And also, apparently, the giant desert centipede might also give you something that is one step about, unfortunately. And then finally, the other scale I was looking at very closely, it's basically the Scoville Scale Heat Units Scale, which is the intensity of different kinds of peppers. And ultimately concentration of capsaicin. And I mean that when the Scoville Heat Units are linear. And if you map them on to basically the intensity of the pain that people report, you can arrive at this logarithmic conversion, we would hypothesize that the Scoville Heat Units actually might map on more or less straightforwardly to the actual brightness of the pain quality that you experienced. But then, if you describe that, basically, how people will rate if you see the ratings of hot sauces, they will rate them as kind of four out of five, or three out of five. And like the 3 out of 5, may have something like 50,000 Scoville units, whereas the 4 out of 5 might have something like 200,000 Total Units, and maybe the 5 out of 5 is like a million or more. So that is kind of the logarithmic to linear conversion. I'm trying to explain.

SPENCER: I see. So the mapping of how people rate them, let's say on a 1 to 5 scale, seems to be going up exponentially in the number of Scoville Heat Units.

ANDRES: That's right, exactly.

SPENCER: I think there are two related concepts that are very easy to kind of mix together. One of them is if you have a human rate pain on like a one to 5 scales or 1 to 10 scale, how do they in practice rate things? And do they sort 1 use a logarithmic scale implicitly where, let's say you're doubling every number or going up by 50% of the number, something like that? The second thing is, is there a sort of exponential relationship between input and output? So for example, let's say you start giving people increasing electric shocks as you're increasing the strength of electric shocks, does the experience of pain go up exponentially? Or is it go up linearly? Do you want to just comment on those two kinds of related ideas that are distinct?

ANDRES: Yeah, this is a very, very central point, in fact, and it is extremely confusing because, there is this thing called the Weber's Law in psychophysics, that basically says that, for you to experience what's called a just noticeable difference in the intensity of input, you need to change the input by a certain proportion, as opposed to by a certain absolute value. So basically, like for you to tell whether, like two objects in different hands that have the same weight or different weights, it's not that they have to be a threshold of like 10 milligrams of difference is more like they have to be a threshold of, you know, 3% of the difference in terms of their weight. Now, some people interpret these as, hey, like, there is a logarithmic in a sense from the version from the intensity of the input to the intensity of the experience. And from that point of view, one could argue that, hey, actually, pain scales are truly linear, is more that kind of the intensity of the stimuli may be exponential, but that doesn't matter, because it translates into like something much more pain at the level of differences in intensity qualia. The logarithmic scales of pleasure and pain make it very different interpretations, in a sense of interpretation is that at least within certain windows, especially window adaptation, so basically, once you have adapted to a particular level of intensity of given stimuli, then differences in proportion in that stimuli do a map on to differences that are equal in proportion in the intensity of the qualia, that you experience. However, for you to tell apart, two qualia bundles, in your experience, that are slightly different intensities, for you to be confident that one of them is actually a higher intensity, it requires to be a proportion higher. In other words, Weber's Law, we think, is not only in a sense about the differences in intensity of stimuli, it might actually also be about differences in intensity of qualia. Because if you're experiencing some extremely powerful pain in your left arm, and extremely far face painting in the right arm, but they're not substantially different in a proportionate way, you will probably just not be very confident that one is stronger than the other. But you could still tell something like the other probably like way more intense, then, you know, something like one paying grade lower than the other.

SPENCER: So seems like this might be rather settleable with a scientific study. Imagine you brought someone in and you gave them different levels of electric shock, obviously, with their consent, and you started with like 1 you know, electric shock, then 2, then 3, and then you ask them, okay, you now get the choice. Do you want to do one unit for three minutes? Or would you rather do three units for one minute? And you could like kind of create an exchange rate between that? Do you think that would help settle this question?

ANDRES: I think there would be good, good information, for sure. It might not fully settle it simply because there are strange effects that something has happened when we aggregate or like, in a sense, I think of aggregating pleasure and pain over time. And I think like, yeah, there's some research about how people even prefer longer duration of pain, as long as the end in relatively lower pain, so that we kind of like create this, I guess, a perceptual illusion.

SPENCER: But like the peak and raw, that kind of thing.

ANDRES: Yeah. I mean, I think it would be useful. And it would be one other data point that I think would kind of like be part of the consilience that fortified basically the case for logarithmic scales.

SPENCER: If you could do any study imaginable like this, you know, that it actually is possible today's technology what study would you want to do to help pin this down?

ANDRES: Yeah, I mean, this definitely goes, I guess, deep into the whole issue of quantifying consciousness. But, I guess I'll bring up an important concept, which is that we're very interested in basically, paradigms identify how intensely conscious a person is with essentially amounts of energy that are yet basically trapped in a particular system and a particular neuroscience paradigm that we're very interested in and we are actively doing research on URI is connectome specific Kermani waves, where you identify what are the resonant modes of your connectome, including your cortex. And then based on your imaging data, you infer what is the most likely weighted sum of harmonic modes. And what's very nice about this is it actually outputs a total amount of energy in your consciousness because each of the harmonics basically has an intrinsic energy which is a function of their frequency. And also you can quantify the amplitude, so basically, by weighting the amplitude by the intrinsic energy of the harmonic, and then adding all of that up, you can get an estimate of in a sense how energetic in a particular state of consciousness is. And when they have done this analysis on, for example, LSD, it's pretty unequivocal that the LSD state of consciousness is simply put a more energetic state of consciousness. Whereas, for example, propofol or basically sedatives give rise to basically sets of processes at lower total energy. Likewise, if you look at the analysis, I was pointing out at the beginning of identifying cascades of activities in neural networks, you can, in a sense, plot the overall energy that is going on in the network at any given point in time. And that would follow a log-normal distribution, I mean, to the extent that energy will be roughly proportional to the number of neurons firing, and expect that if you basically point this type of analysis, to whatever is going on in a nervous system, when somebody is experiencing a cluster headache, I'd expect that you would actually see outstanding levels of energy in a highly dissonant form. And if we were to plot that, basically, I would expect, even within a given cluster headache, you will also have a lot of normal distribution, and basically how intense these spikes are, as also something that we have done. But in from interviews that people say that even within a bad cluster headache, they're fractions of seconds that are even worse, especially worse. So even within a very bad episode, we might expect to still see basically a log-normal distribution in the energy spikes of the activity of the network. And I think that would probably maybe settle these matters.

SPENCER: So I have so many questions about what you just said, I really want to understand what you said because I think it's actually there's so much just packed in there. So you said you were to connect to home if I take that as meaning sort of the connections between all the different neurons in your brain.

ANDRES: It's much more a macroscopic level description. So the way you gather this data is with that fusion, tensor imaging, where you model how water moves within your brain, essentially, and the directions along which water can move more freely, essentially map onto regions of the brain that have white matter tracks. Basically, these are like very long axonal connections, and they tend to be bundled together in this, I guess, like bottles of cables. But definitely, if you Google connectome pictures or diffusion tensor imaging, you get these really awesome pictures of basically the inner structure of the brain, and how even parts of the cortex are connected to each other with these very big cables. So the innocence and basically the connectome plus the cortex, you know, you also need to essentially map on the actual kind of like the way in which the cortex is folded for a particular brain and does like slight differences between people. And there's this additional issue that waves travel faster in white matter tracks, as opposed to grey matter, grey matter would be the cortex. And once you take those factors into account, what you can do is essentially model, what are the possible standing wave patterns that you get as a result of kind of this entire system. And the very nice thing is that you get a discrete number of possible standing wave patterns, which are, you could describe as the harmonics of the brain, or the resonant modes of the brain.

SPENCER: Which are the standing wave patterns. This is like the idea that if you take a violin string, and you pluck it, there are only certain ways that it can vibrate because it's kind of being held on each end. And you can have it sort of divided one time or two times, and so on. And these are the different sending waves.

ANDRES: That's exactly correct. And a very beautiful way of seeing a physical demonstration of these is with what's called the Fadli plates, which are these metallic plates usually shaped like a square, but they can be shaped like anything. And in a sense, they will have the discrete number of resonant modes, which is basically all the ways in which kind of these mechanical waves are going up and down in feet, an integer number of times within that shape.

SPENCER: I think I don't really understand what's resonating in the brain, though.

ANDRES: Yeah, in this case, it would be basically electromagnetic waves. So everything is basically a signal propagating through an axon is kind of crazy because it's not that in a sense, anything is like truly kind of like moving along there is more than that, ion channels are like getting open along the way. And there's a change of electric potential that happens everywhere on yourselves get open, but as a consequence, you get these flow in the electric field. So ultimately, it is kind of like an emergent electric field flow that you get. And the idea is that there's basically an integer number of ways in which that can stay in a standing wave pattern.

SPENCER: But you do fit a lot of really high-frequency waves in there is it just that the frequencies are not that high. So that's why you don't get that many different wave patterns.

ANDRES: You get a lot, I mean, like, you get, like 1000s of harmonics. But on most states of consciousness, a lot of the energies are basically concentrated in the lower frequency.

SPENCER: Now, when you talk about the energy being concentrated, are you essentially saying you do a Fourier transform so that you go from thinking about this thing, oscillating a time to instead saying, Let me chart the strength of these different frequencies? And then if you look at those, essentially, the amplitude different frequencies, that is kind of proportional to, or at least related to the energy in that frequency that we're talking about?

ANDRES: Yes, yes. So, this is really cool. Because when you tell a person, you're able to say something about the brain, at the level of very high frequencies and something like 10 hertz or higher, using fMRI, it's going to be very puzzling, right? Because people tend to say, it's true, EEG is very high temporal resolution, but very low spatial resolution. An fMRI is the opposite, very high spatial resolution with low temporal resolution. So how is it possible that you can infer the way that some of these pretty fast harmonics, just based on what's ultimately very low, a very coarse-grained temporal resolution pictures of the brain? And the answer is that you use a probabilistic model, in a sense, you're looking at the data with very high spatial resolution. And then your model is telling you what is the most unlikely combination of harmonics that over time would result in this particular pattern that you're looking at. So it would be sort of more or less equivalent to if you take a violin string or a guitar string, and you pluck it, but you don't have kind of a high-speed camera. But maybe what you do have is kind of a service topic light, so that you can cast a shadow of that string, and take a picture, you know, several times per second. And you can then like, look at the shape of that shadow over several seconds, and infer what is the combination of standing wave patterns that are giving rise to these particular states.

SPENCER: So it's a great way to think about this as essentially like a graph structure in a mathematical sense. You've got nodes and edges, and then you have waves propagating between the nodes and then kind of the frequency analysis on this entire graph structure.

ANDRES: That's exactly the way. Yes.

SPENCER: Got it. I see. And then I think you're saying that the amount of energy is somehow related to the qualia, we experience or in other words, what it feels like to be us when we're in that brain state to talk about that relationship.

ANDRES: Yeah. So this is what they're finding with this particular paradigm, that if you analyze the brain of psychedelics, you will see that not only in general, you have a higher amplitude for these harmonics, but you also have a higher amplitude, especially on the high frequencies. And in fact, actually, you get a little bit of a deep amplitude in the very low-frequency harmonics. This is true for psilocybin, for DMT, for LSD, and even for ketamine, which is kind of surprising, right, because ketamine, in very high doses, you know, definitely a kind of sedative, it puts it out, but in some anesthetic doses, it actually enhances consciousness is getting densify consciousness in several ways. And that is what they show with the scans to this analysis, as opposed to propofol or other kinds of like more classic anesthetics that are described as kind of a real turning down the volume of consciousness. If you do a connection, specific harmonic wave analysis, you will see that they radically decrease the amount of energy in the high-frequency harmonics, they may increase the energy a little bit in the very low-frequency harmonics, but the aggregate total energy bills substantially down.

SPENCER: Okay, so how accurate are these measurements? So you mentioned that you have to make some assumptions because the tools have limits to their resolution and speed. So do we know that these frequencies are what the tools tell us? Or how much guesswork is involved here? I guess I'm asking.

ANDRES: That's a great question. I'm not sure it is quite interesting that you can use the disconnect harmonics as features to predict a particular state. And they seem to be really good at that. So, for example, if you look at the core occurrence patterns, particular harmonics on a psychedelic versus on a sober set of consciousness as you can very easily say that, okay, this particular fMRI must be somebody on LSD or something like that, because these harmonics usually get on come together normally. So that's another general finding. It's called enhancing the repertoire. That is a sense, there's kind of a very fine grain pattern of correlations between even harmonics in resting state and I guess like normal states of consciousness, that usually kind of get extremely shuffled around on psychedelics like completely new combinations of harmonics become possible on psychedelics. So, there's definitely a good number of inbuilt assumptions on the model. And ultimately, I think some of them might probably be violated in some circumstances, like, for example, assumptions of linearity, or like superposition principle between these harmonics, I very much expect that on high energies, basically, superpositions will give rise to nonlinearities. And like, that's something that this paradigm is currently not modeling. Although there's like some conversation, some have extended, so that also takes into account nonlinearities. But by a large, I've been very impressed with it. At the very least, it's, you know, descriptive and predictive power, that if you use these harmonics as features, it can essentially tell you what state of consciousness you're in, which is really impressive, I think. And it's just such a radically different way of looking at the brain, then something like examining how intensely your amygdala is activated, or something like that kind of dysfunctional localization paradigm, which I think it's also very useful. It's another lens, but it doesn't give you this very nice property of like energy, it doesn't tell you how energetic a state of consciousness is, whereas connected harmonics, they do. And they seem to be consistent with subjective reports of the intensity of the experience.

SPENCER: This reminds me a lot of spectroscopy. Like if you want to know what a chemical is, you can look at the sort of spectral lines of electromagnetic radiation coming off of it. And you can decide, oh, is this hydrogen, this is helium, etc? And it seems like there's an analogy here, which is, you're looking at the spectral lines of the electrical signals going to the brain and saying, is this person on LSD or they're in ketamine? Or, you know, did they just see something they love? Etc?

ANDRES: Yes, yes, that's right.

SPENCER: And so I think you're also arguing that you can basically you can do these brain scans, run some kind of algorithm over, get these spectrums. And then they have good predictive accuracy at predicting what state someone's in, but how much should we read into this? Like, okay, so we these are things that differ for different brain states. But does that mean that they're kind of explaining something fundamental about the brain states? Or are they just sort of averaging over a bunch of properties, and they happen to allow you to distinguish the brain states?

ANDRES: Yeah, that could be possible. So we definitely need to continue investigating this paradigm. It is just I would say very promising, given the data that we have. But if you're very well turned out of it, it's kind of a four dimensionality reduction technique. And like, there's like some other ways of looking at it. But the thing that's, you know, it's very, very attractive from kind of the accomplishing something desirable with this analysis. It gives you an energy range, and a sense of repertoire, that basically, a state of consciousness, thought of as a weighted sum of the selected harmonics can be described as either normal, or I guess, an everyday sort of consciousness, or something exotic, just based on like, basically, the patterns of correlations between these harmonics. And, yeah, that's something that it's a lot harder to do, I think, on other paradigms.

[promo]

SPENCER: So this might be a good time to talk about valence. Do you want to kind of explain your theory of valence and how that relates?

ANDRES: Yeah, definitely. So this is I mean, one of the main things that I'm interested in, it's one of the main research topics that you're right about. And it connects with specific connectors, specific harmonic waves, and also logarithmic scales of pleasure and pain. And overall, the main claim is that whether a set of consciousness feels good or bad, ultimately depends on the structure of the state of consciousness, which is a very different way of looking at it than traditional accounts of like, what pleasure and pain are, and I mean, like I can mention a few common paradigms. So, first of all, you have kind of this functional localization paradigm where like people might say, hey, pleasure is fundamentally just activation of the pleasure centers, you know, but in some sense, that's just an empirical observation that like, Okay, this part of the brain seems to be very correlated with kind of Valence states of consciousness. But just because it calls something a pleasure center, it doesn't mean you actually have an explanation for why it is a pleasure center. And so we think of that as much more of a hint than, you know, kind of like a final explanation. Likewise, with, for example, something like neurotransmitters, that a lot of people say something like, Hey, let's see, or Tony needs a happiness neurotransmitter, or dopamine or endogenous opioids. And, you know, obviously, they are part of the explanation, and they're going to be a piece of the puzzle. But empirically, you know, you can inject opioids in many parts of the brain, and not have a photonic response, you actually need to inject opioids in very specific parts of the brain before you have like any kind of actual fibrotic response. And there's like some parts of the brain where injecting opiates actually feels this for it. So kind of like just an explanation based on, you know, his particular molecules is yeah, basically, very incomplete, at best. And then, you know, there's this entire way of looking at the brain, which we might call kind of the computational theory of consciousness or computational theory of mind, where they might identify things such as the quality of pain or pleasure with events that happen in particular algorithms. And these might be something like associating pleasure with positive reinforcement; in reinforcement learning algorithm, or, you know, pain with a negative reinforcement signal system.

SPENCER: So this would be like if you build an algorithm that's trying to optimize for some goal, and it gets points every time it gets closer to that goal, those awarding points in the algorithm would be linked to pleasure in that sense.

ANDRES: Yes, exactly. And even implicitly, in something like the backpropagation algorithm, in artificial neural networks, this idea that you are rewarding the network, whenever it's making a correct prediction, by, in a sense, strengthening the connections that give rise to that prediction, or you're punishing a network, when it makes a run prediction, by likes, kind of propagating, like slight adjustments in the weights to biases towards making the correct prediction later. And in our experience, with a lot of very smart people, indefinitely people in our shared communities of interest. That is the general way in which they tend to think of Valence is kind of like an emergent effect out of like, something that can reinforcement algorithm, or yeah, basically something that it's happening at the algorithmic level.

SPENCER: I struggled to understand how reinforcement learning algorithms like the kind of points rewarded for optimizing could map into valence, because of the sort of scale issue, right? Like, if you think about minimizing an error function, you can also think of that as doing the reverse maximizing the negative of an error function. Or, you know, you could always cut the number of points, you get it in half, and it's still gonna optimize the same function. So it seems to me there's this weird scaling issue there.

ANDRES: I agree. I agree. And I think that's a very, very big point. I mean, the way I've heard people kind of wrestled with that particular point is something like that what matters is the relative states. So like, maybe the absolute scale doesn't matter but is the effect it has on the system. And whether it's, for example, strengthening or weakening connections in an artificial neural network, which kind of like something that happens at the relative level, and you need to take into account the entire system to arrive at the meaning of a particular intervention.

SPENCER: Even that seems like a problem because any monotonic, increasing mapping done to the objective function shouldn't change the answer, right? So like, you could even distort the relative strengths of the point system. And that seems like you'd still be optimizing for exactly the same thing. Yet, from this point of view, the rewards would be totally different. So yeah, I guess I don't really get how to resolve that. I mean, it seems like a tricky question to me.

ANDRES: Yeah. Yeah. I mean, in the end, like the way that we resolve it, is by saying that, yeah, Valence is not something that happens at the algorithmic level is something that happens at the implementation level, basically, that it is not so much that states that are operations to a brain that are reinforcing device the pleasure are much more than in a sense, pleasure is being recruited in order to instantiate a reinforcement learning algorithm. And you can increase it or perhaps recruit other properties other than pleasure and pain, but that they have some qualities that make them very, very, in a sense, like efficient and at a very, very high level. There is this concept called the symmetry theory of homeostatic regulation. And the overall idea here is that if you know, you're kind of like computing the harmonics of the system, even, for example, the harmonics of a, you know, a bacteria or something like that, that, by examining those harmonics, you can tell whether the bacteria is damaged or not. Because whenever there are symmetry-breaking operations that happen to a particular system, you will essentially see that very clearly as dissonance in the harmonics that arise. So, just to give you an example, like, if you have a sphere, and you found it here, you metallics here, and you get it and you hear it is going to have a particular tone. But if you make an indentation in the sphere, you will hear kind of spurious sounds that, like its harmonic signature will be in a sense, damaged, if you deviate from symmetry in any way. And the idea is that the reason why we're trying to optimize kind of these symmetrical configurations is that they are the way in which we regulate homeostasis, that in order to kind of like diagnose problems that you may have, in your system, in your biological system and your nervous system, kind of having these very quick ways of testing for the symmetries of the system, is a way in which you can very quickly and massively parallel way detect and diagnose problems. I'm not sure if I said it very clearly. That's roughly the idea.

SPENCER: You're saying if a system is symmetrical, it can be a way of showing that everything's kind of working away. That's expected because then any deviation from symmetry can be a quick measurement of something that went wrong. Is that the idea?

ANDRES: Yeah, that's right.

SPENCER: What if the harmonics correspond to though in the kind of systems you're referring to.

ANDRES: So whenever there is an actual symmetry, in a system, there will be a corresponding set of harmonics for that. So for example, if you have a guitar that is like, bilaterally, symmetrical, there's going to be a harmonic that basically goes from left to right, and it's oscillating, left to right. And like, for example, in the brain, the lowest frequency for like, one specific harmonic wave is one that basically has in one, at one moment, you have all of the left hemispheres activated, and the right hemisphere deactivated, and then they flip flop, you get the right hemisphere activated, and, and vice versa. So if you have any kind of asymmetry, imperfection, or deviation from what a train that harmonics, that will show up as that harmonic basically being slightly different than what it would be otherwise. And you can use that as kind of a very quick measure of, hey, hold on, there's been some damage here. And in a sense, honing in on that damage by analyzing the particular patterns of the harmonics.

SPENCER: So is the idea that if you had a guitar, and you strummed it, it would make relatively pure tones, not perfectly pure, but relatively pure. And then if you were to damage it by, let's say, making a nick on one side of it, that that would actually make kind of more of these kinds of unfair tones. So your frequency spectrum will get more kind of blurred out?

ANDRES: Yeah, yeah.

SPENCER: Got it. And you mentioned that maybe I misunderstood you, but I thought you mentioned sort of this harmonics in a system of bacteria. Is that right?

ANDRES: Oh, yeah, I guess what I'm saying is that this principle is homeostatic. Sort of symmetry theory or from a static regulation may apply both to nervous systems, but also potentially even simpler organisms. Like even something like bacteria.

SPENCER: What is the wave in the bacteria?

ANDRES: Well, it would be like, both electric and mechanical waves that propagate across the membrane.

SPENCER: Also, like pressure waves, like sound waves or electric electromagnetic waves. Yeah. Okay. So, you have this idea that maybe symmetry is a kind of measure of things like working properly. So then how does this connect to valence?

ANDRES: Yeah, okay. So, a couple more things. The other thing is, for the most part, you will find that these symmetrical states are also kind of like the lower energy configurations of the system. So, for example, if you look at a soap bubble, you know, it will tend to organize as a perfect sphere. And the reason is that any other configuration has too much potential energy, which will be translated into kinetic energy. And because of its environment, you know, that it has like the air around it, in essence, is going to be radiating that energy. And the only configuration where like, cannot radiate any more energy is going to be the perfectly symmetrical one, which is the sphere. So in essence, it's kind of like his general tendency of systems to kind of like settle on low energy configurations, and also there is this tendency for low energy configurations to be symmetrical. So the idea is that at a fundamental physical level, what we are doing kind of constantly is trying to lower our overall energy, and the impediments to doing that, is ultimately the source of suffering, where you're kind of like trying to settle into this beautiful combination of, you know, low energy configurations, but you have something like an irritating sound, or like some chronic pain, or something, you can think of that as an irregular signal that is constantly pulling and pushing you away from these, like lowest energy configuration. And I guess like if you meditate a lot, and you achieve kind of these very tranquil states, you can kind of introspect on these subjective feelings of how, basically, you're utilizing energy in order to identify small deviations from perfect symmetry. But, you know, left and right, up and bottom, front, and back. And that the things that bug you, in a sense, are things that are constantly kind of like trying to push you away from those harmonics, and that the work that we do, in order to get back to that beautiful symmetrical settled state, is usually competition with useful work. So kind of the idea is that we evolved to be configured in such a way that our efforts to basically go to this very beautiful from your static state does actual useful computational work, I'm reminded of this idea by, for example, Richard Feynman, being confused is a very unpleasant state of consciousness. And trying to understand something requires a lot of confusion. So in a sense, like for you to do computationally useful work, you need to be suddenly up, in a sense, you cannot be in the perfectly symmetrical, meditative, highly settled state, you have to be kind of like jeter away from it, you could still be a pretty high valence state on the whole, like, it could very well be that some parts of your nervous system are settled and are like, it's very beautiful from your set of states. And some parts are not that in aggregate, you still feel good, you know, you could be in a comfortable chair, sipping hot chocolate, while working on a physics problem. Overall, it could be a pretty pleasant experience, but the actual confusion of trying to solve the physics problem required cannot be being unsettled, not actually being in the lowest energy configuration.

SPENCER: So struggled a little bit to know what symmetry means here. You know, we talked about these electrical impulses going to the brain and, and having some kind of frequency spectrum. And you talked about, you know, imagined left and right hemisphere. And you can imagine you know, impulse waves going back and forth in a kind of symmetrical fashion. But like, how does the symmetry idea map onto this idea of kind of frequency spectrum in the brain?

ANDRES: Yeah, sure. So there are a couple of things, one is select symmetry, broadly speaking, we talked about it as invariance of transformation. And the particular kind of symmetry that happens on harmonics is symmetry over time, but it also manifests as a spatial symmetry. And there is a crazy duality between the two. So I can just share an example that like, if you take a psychedelic, you will notice kind of like all of these crazy frequencies going on, simultaneously. But if you can meditate and harmonize them, and make them very pleasant, you will notice also that the perceptual illusions and hallucinations are also changed character. So for example, for you to experience a perfectly symmetrical hexagonal grid in your visual field, which would be kind of these tessellations. So that has like spatial symmetry, for you to create is that you require a lot of harmonics that are happening, you know, in the temporal domain, to be synchronized to such that they come together at exactly the same time in order to paint to these perfectly symmetrical features. So there is a duality between temporal symmetry and spatial symmetry in resonance systems. And in some sense, there's kind of, yeah, really equivalent. I mean, like, there's the, you can think of it as kind of a four-year transform that properties in one domain map to properties in the other domain. And here, generally speaking, when you feel for example, balanced, spatial, like you feel that there's like not much more weight in there in your left or right part of your body, that actually also corresponds to a particular pattern of repeating an activity over time, that is completely constant, and it's not deviating from that constancy. Just to clarify, like with a technically, you have a metallic plate, and you sprinkle sand or salt on top of it, and then you play a pure tone, and in that pure tone is close enough to the frequency of a resonant mode of the plate. That plates will basically get that resonant mode to be energized. So like, you're kind of taking some of the energy of the pure tone and putting it into the resonant mode of that particular plate. And that the shape that appears in the sand that you put on the plate will be asymmetrical configuration. So it could be kind of like his flower pattern or, you know, this tree. I guess the flower pattern or like Mandela pattern, like a lot of mandala patterns emerge. And because the shape is symmetrical, like let's say you pick a square, those patterns will basically also have the symmetries of the shape that you reduce. So you will have something like they will be bilaterally symmetrical, and also they will be symmetrical across the diagonal. Now, if you play two pure tones at once, one that hits a resonant mode of the fleet, and another one that may be a slightly different one. And then you look at the emergence standing wave pattern of simultaneously activating these two resonant modes, for the most part, their shapes will actually be unstable. In other words, like, rather than just having this kind of like very pretty symmetrical standing wave pattern emerge, that is static over time, when you combine harmonics together, you can get something that for example, is rotating, or something that is kind of like unfolding or something that is kind of expanding and contracting. And if you add more harmonics, especially you're careless about how you do it, if you're picking random harmonics, the standing wave patterns will be very complex, and will be very unpredictable. It's kind of like, there's a lot of these frequencies, that because they don't fit an integer number of times with each other, basically, when the pattern repeats is going to be the minimum common multiple of all of them, right. So if you had a bunch of random frequencies, you will get a standing wave pattern that really doesn't repeat in a very, very long time. And in that sense, it really likes kind of spatial and temporal symmetry. But if you are very judicious, in which particular frequencies you choose to simulate the plate with, you can get something that is also stable, that you can actually get symmetrical standing wave patterns, even though is the weighted sum of multiple harmonics.

SPENCER: Got it. Okay, so do we have enough background now to start linking this more directly to pleasure and pain?

ANDRES: Yeah. So, in a sense, our overall paradigm, which is something that we're currently investigating, both in neuroscience and in neuro-technology, is this idea that yeah, basically, spatial symmetry will map on to particular temporal symmetry as well. And that, you know, harmony in sound match on to temporal symmetry. And when you simulate one of these, you know, harmonic resonance systems, with these symmetrical patterns, you will also get a spatially symmetrical configuration. So, the net effect of this is that you can, in a sense, characterize the hedonic fingerprint, or the hedonic signature of a state by looking at, in what ways over time, the system is resonating in ways that are A consonant, B dissonant, or C noisy, which is neither consonant nor dissonant. And very similar to basically in music theory, if you want to quantify basically how dissonant a combination of notes in the piano are, what you do is that you take the spectrum of each of those notes. And each of those notes will generally basically have a fundamental frequency, and then its harmonics. So basically, piano notes, maybe something that you play, and the bass frequency is 200 hertz, but it will also have a weighted sum of 400, Hertz, 600 800, and so on. And when you combine two piano notes together, what you will have to do is look at, in what ways the harmonics of both piano notes are interfering with each other in order to generate dissonance. And it turns out that basically, when you have two pure tones that are very close together, let's say they're like within 20 hertz apart, simultaneously in high energies, that causes these emergent beat patterns. And that's basically a way of breaking symmetry over time, and that is a dissonant configuration. And empirically, this is known since the time of Helmholtz, that the overall dissonance of a combination of musical instruments is going to be the total sum of the dissonance between each pair of harmonics that is present in the spectrum. And of course, like most harmonics are not interfering with each other because they're pretty far in the spectrum. But whenever you have kind of like, lots of crowded harmonics, pretty close together, that is going to sound like very abrasive, very rough and very, very unpleasant. And in a sense, like the art of composing music, is the art of balancing enough complexity so that you don't get bored. And in theory, boredom is a type of business as well. But yeah, basically, you're avoiding boredom, while also making sure that you produce as little dissonance as possible. And musical keys, for example, are substance, you know, of a musical scale, such that the pairwise average dissonance between the notes in the key is generally minimized. And how much is minimized, will actually tell you the overall hedonic quality of the key. So like basically major fees, on average, the pairwise dissonance of the nodes in a major key is much lower than the average pairwise distance of the nodes in a minor key. And when we suspect, what we're trying to test is whether some analysis like these can be applied to nerve resistance, that in a sense, when you're in a good state of consciousness, especially, let's say you're in that it gets tricky, like you've been feeling good for, for several days, that, in a sense, we are anticipating that the reason why you're feeling good is that your brain is selecting a repertoire of states that are built upon a subset of the harmonics that you have available to you. Such that the average pairwise dissonance between the harmonics that you're experiencing is relatively low. In other words, somebody was like, very happy is actually leaving, in a what Mike Johnson, you know, positive effect, T signature of experience, that the building blocks of their experience are harmonized with one another. Whereas may be experiencing a lot of stress and chronic pain, depression, and so on. We predicted that the repertoire of states would be drawn from a kind of like minor key of your nervous system, that the pairwise average pairwise dissonance between the harmonics that you're experiencing, is on the phone pretty decent.

SPENCER: So I think the one thing that confuses me a little bit about this is that if you think about music, a pure tone, like a pure sound wave, well, it's not extremely unpleasant also is not the most pleasant, right? There's something a little bit annoying about it. I'm wondering, Is that relevant here?

ANDRES: Yeah, definitely. So that's why we have to bring up essentially the concept of boredom, boredom, we think of it as an unpleasant experience that is different from the stimuli to which you're bored. So when you're bored, it's kind of like this oppressive, kind of like, the cloud of negativity, in your experience. And ultimately, we think like, you know, boredom will come down to some kind of neuronal dissonance as well, you know, because you can ultimately cure boredom with all sorts of things, you know, like, drugs, or simulation or for medication. So when you hear a pure tone, over and over, you get bored. And that is what makes it unpleasant.

SPENCER: To sing, it would be the most pleasant because it's like the most symmetrical, but there's this extra kind of boredom thing that makes it kind of annoying over time.

ANDRES: More or less. Yes, I mean, that is the first approximation, I would add one additional caveat, which is that with just a pure tone, basically, you're constrained on the amount of energy that you can arrive at, because, I mean, you can increase the volume of that pure tone, but at some point, that's going to basically damage your ears, and also you can, you know, represent just a pure tone to the same level of energy as you can represent a very, very complex experience. So, in a sense, if you actually want to maximize valence, you need to simultaneously take into account what is the overall harmony or consonants of a state and what is the overall energy of the state. And if you only have one pure tone, you're not going to be able to max out the energy.

SPENCER: Does energy depends also on the amplitude like higher amplitude means higher energy and on the frequency with higher frequency means higher energy?

ANDRES: Yeah.

SPENCER: So with that suggests that a very high-frequency tone would be more pleasurable than a low-frequency tone.

ANDRES: Not necessarily, I mean, again, this has to do with how much energy is like spreading to other harmonics.

SPENCER: Let's say it appears or not, so it has no harmonics.

ANDRES: So, I mean, I can give you an example which is some people describe like, very, very advanced, like stages of samadhi as samadhi, which would be a kind of very, very deep meditative states as having a very, very clean, extremely high frequency. And that might be Yeah, something akin to a kind of like a very, very high frequency, pure tone, and that it's actually extremely high valence because there's nothing interfering with it. It's kind of on its own, it doesn't have any dissonance.

[promo]

SPENCER: So let's suppose someone gets pricked with something sharp, right? And then like they get a kind of sharp experience of sharp pain. What would you expect to be happening kind of in the brain that would be at least in theory, measurable that shows that that's a painful state instead of a pleasurable state?

ANDRES: Yeah, we roughly expect that there's going to be a lot of dissonances, and there's like how you quantify the incidences, you know, a rabbit hole, but uh, that I mean, any kind of like negative stimuli will basically jeter you out of your native connecting harmonics. And that is going to be an unpleasant state. So, when an example is a fibromyalgia, which is one of the most painful conditions, especially for some people, and this way of modeling fibromyalgia that is called explosive synchronization, where basically, very similar to the log-normal distribution of, you know, the neural activity I was talking about earlier, except that this is kind of an extreme case, you know, where like, very, very light stimulation in let's say, the scene of somebody with fibromyalgia can feel like, you know, these like very strong cascade of activity all of a sudden happen, it's kind of like it the nervous system is extremely sensitive, in fibromyalgia. And the ways some scientists model this with a paradigm of explosive superposition, where they look at, okay, what is the probability that you will have like very large assemblies of neurons, suddenly, that start seizing, and in our model, okay, like seizing, you know, kind of like spontaneously in your peripheral nervous system, without feedback from your central nervous system will basically be a super strong, kind of like metronome frequency that is being injected into your central nervous system in a way that doesn't allow you to adjust for it. And because the, you know, whatever, random resident modes of whatever patch, or assembly of neurons in your peripheral nervous system got activated, is very unlikely to actually fit together well, with a native harmonics of your brain, those very strong signals from outside will in a sense, be in a very, very dissonant relationship with the internal native harmonics of your brain. And we explain basically, in our paradigm, we explain why Fibromyalgia is so painful, it's for that reason, that is basically creating these extremely misaligned, very, very high energy interface between your peripheral nervous system and your central nervous system without getting time to adjust. On the other hand, if for example, you were to kind of have explosive synchronization in the rest of your nervous system, but if it could be harmonized with the native harmonics of your brain, then those may actually be extremely beautiful kind of ecstatic states of consciousness and you one piece of evidence in this direction is, like what happens if you look at the EEG of people in deep Janus, there's something very, very peculiar happens, which is that many of those states actually show very, very similar signatures to epilepsy, except that the harmonic structure of the EEG agenesis difference, because let's say in epilepsy, a very difficult pattern is that you get kind of this very powerful, like four-hertz side, me like the full picture, you can find it in symmetry theory, valence 2020 presentation, including computing, which is a presentation we gave to Robin Carhartt Harris and other people in his team. And yeah, it was very well received. And I think it does a much better job at explaining the overall picture of the symmetry theory of valence and all the evidence that suggests it's the case than I can possibly do right now.

SPENCER: Cool. So are there any other things you want to say about your theory before we wrap up?

ANDRES: Yeah, I mean, like, the very nice thing is that, if it pans out, and you know, something that we are actively researching, we are currently analyzing datasets of people on MDMA, and 5-MeO-DMT. Now that we conduct those studies ourselves, basically, we're collaborating with, yeah, other like neuroscience labs, that have the permission to basically conduct the studies, essentially, so many, if this pans out, it might be, you know, the first way of actually quantifying valence from first principles in a way that generalizes to other species. So that means like, yeah, very important, from the point of view of EA as well, that it may very well turn out to be that, you know, being a, a dung beetle is extremely awesome, because the native harmonics of the nervous system or dung beetle, like naturally tend to send sheets, you know, major effective key signatures, as opposed to like minor effective key signatures. Like, you know, maybe being a shark, it's inherently very unpleasant, that could also be an outcome of this. But importantly, it will allow us to, in a sense, like do prioritization in a completely new way that it will have an objective understanding of, okay, how much energy does this negative valence state is trapping, you know, the, how intense is this negative valence state, and we can study that across humans and also across other species, we will be able to identify, you know, like maybe the 1% of like negative states of consciousness, then if we're able to prevent or treat, that we will be able to perhaps get rid of something like 90% of suffering because most of the suffering is concentrated in those very, very unfortunate, super intense states of consciousness. And to the extent that this is a blind spot in effective altruism, I think it's yeah, it's very important to figure it out.

SPENCER: So I'm curious how you would respond to someone who might be skeptical of your theory, like, what would you want to say to them?

ANDRES: I guess it really depends on the specific kind of skepticism, there is, under the whole, that kind of skepticism that comes simply from different theories of consciousness, like, if one is a functionalist, as opposed to, let's say, a physicalist. One, we criticize the idea that there is actually kind of like a matter of fact, about like how intense an experience is or something like that, I do not like to address that type of criticism, I would usually have to go into the philosophy of mind, kind of like very, very deep into it. If that is kind of the blockage. In the case of, you know, one of our listeners, I would recommend watching a video I made very recently called Digital Minds, can digital computers ever wake up, where I basically outline the very high-level picture of why we think physicalism is much more promising than functionalism or computational theory of consciousness. There are also other criticisms about like, is it even possible to kind of like report to the intensity of a state of consciousness. And, you know, for that kind of criticism, I would say, you know, actually, just firsthand experience probably matters, like experiencing different ways of hot sauces, and kind of like, identifying you just objectively convincing yourself that like, oh, gosh, yeah, this is like, differences in levels of energy, not on like, like, what it's like to basically experience electroshocks with different voltages or experiencing heat at different like temperatures that like there's also kind of like this notion of energy in like in purely internal hedonic states. And, yeah, that's something that it's best to be convinced on a first-person kind of basis. If not, then yeah, perhaps like reading the whole, logarithmic scales of pleasure and pain article for a lot of yeah, the first thing that counts and all the conversion pieces of evidence might be helpful.

SPENCER: Is there one or two things you want to point to that you think, strongly links, this idea of pleasure and symmetry, in case someone doesn't is not fully convinced by the connection between the two?

ANDRES: Oh, definitely. So there's this notion, for example of heart coherence, which is, to what extent your heart is synchronized with your breathing, and your galvanic skin response and other biorhythms of the sort. And it turns out that basically, the more harmonious that relationship is, meaning, they don't have to be synchronized on a one-to-one basis, it's very difficult to have like, one breath for every heartbeat, it's more kind of like, for every six heartbeats, you do a breath, when you enter into those states, those tend to be extremely pleasant. And if you analyze data biorhythms of people in Janus, like very pleasant absorption states of meditation, you will see that they experience very, very high levels of coherence naturally. On the other hand, for example, when you feel a lot of anxiety, or kind of like a feeling of unsettled or irritated, basically, yeah, heart coherence is significantly down. So you can think of this as like, one reason why, you know, heart disease has like, much bigger implications than just, you know, hey, as your life expectancy goes down, is well, no, it actually has all kinds of like, negative mood effects. And we can think of these as, hey, like, the metronome of the heart is kind of like failing to produce this very regular and important metronome that allows a lot of pieces of your nervous system to basically tune to each other. And when that is failing, a lot of the other pieces of the nervous system essentially, kind of get out of tune. And that itself produces a lot of unpleasant dissonances. It's kind of like missing the conductor in an orchestra. That will be Yeah, I guess like another. Yeah, the key piece of evidence that like this happens, even at the level of by evidence.

SPENCER: Awesome. Thanks for coming on. This was a really interesting conversation.

ANDRES: Yeah, definitely. That's fun. Thank you so much for having me.

[outro]

Click here to return to the list of all episodes.


Subscribe

Sign up to receive one helpful idea and one brand-new podcast episode each week!

Subscribe via RSS or through one of the major podcast platforms:


Credits

Host / Director
Spencer Greenberg

Producer
Josh Castle

Audio Engineer
Ryan Kessler

Factotum
Uri Bram

Transcriptionist
Janaisa Baril

Music
Lee Rosevere
Josh Woodward
Broke for Free
zapsplat.com
wowamusic
Quiet Music for Tiny Robots

Affiliates
Please note that Clearer Thinking , Mind Ease , and UpLift are all affiliated with this podcast.