Soul searching: a physical theory for the metaphysical ‘soul’
Izi Stoll is a neuroscientist on the front lines of pinning down the physical components of human consciousness.
When we talk about consciousness, most of us don’t really know what it is we’re describing. Except, that is, for Izi Stoll, director of the Western Institute for Advanced Study. After a 17-year career in laboratory neuroscience, she has been dedicating her recent efforts to exploring consciousness’ physical basis in the brain.
Stoll’s new book, What We Are: The Physical Basis of Consciousness, uses science in an attempt to explain an age-old question: What gives rise to the conscious experience we all seem to have as humans? To shed light on this mystery, Stoll draws from neuroscience, information theory, quantum physics, and even geometry, postulating that consciousness arises from a higher-dimensional space produced by our brain activity.
We talked to Stoll about her theory and what it means for our conceptions of free will, the soul, and other big ideas science is only beginning to touch.
What made you interested in finding a physical basis of consciousness?
When I was a kid, I’d just think about why we exist at all and how we can even think about the fact that we do. We’re not just aware; we’re also aware of being aware. Then, I realized you can actually study the brain and the mind. So, when I went to college, I studied biology and psychology. I see neuroscience as the intersection between the two — how the biology of the brain gives rise to psychological experience.
How do you define consciousness?
I go with Ned Block’s definition. The way I would describe it is in two parts. There’s this streaming, cohesive perceptual experience that we all have — the combination of sights, sounds, smells, pain, temperature, pressure, everything we can sense, put together into one qualitative experience. The second part of consciousness is the feeling that there’s a self having that experience.
What are the dominant theories right now, and what are the problems with them?
There’s the Penrose(-Hameroff) model, where it’s the microtubules inside cells that are exerting quantum effects, and the problem with that is just that every cell in the body has microtubules. It has to be something unique to the way the brain operates. Then, there’s the idea that consciousness comes from a disturbance in the electromagnetic field, and we know the currency of neural networks is electricity, but in this theory, any AM/FM radio should be conscious. So, again, there has to be something unique about neural networks.
And then there are those mystical hypotheses, for example, that there’s some particle that exists in the brain that doesn’t exist anywhere else in the universe, like the corticon, which has unique properties and is responsible for generating consciousness. Hypothesizing a particle that doesn’t play any other role in our physical world is problematic, because that makes it really hard to test. And scientific theories have to be testable.
In general, the most respected theory out there is integrated information theory. The idea is intuitive: Consciousness is integration of all information in the brain. The thing is, all your different senses collect data from the world, and these data are processed in different parts of your brain. There’s no one part of the brain where all the data comes together. Even if there were, that neural activity would still not be the same as mental experience. What I’m saying is, consciousness is not reducible to brain activity, though consciousness and brain activity are certainly connected. The framework of integrated information theory is useful in really naming the aspects of consciousness. But it doesn’t explain how neural networks actually give rise to consciousness. It just asserts that they do.
My approach was thinking, “OK, what is integration?” And so, I came up with a geometrical proof showing what integration is and what it does. It shows that any (n)-dimensional shape that is changing over time must produce an (n+1)-dimensional shape. And the brain, as a three-dimensional structure that is changing over time, must produce a higher spatial dimension.
Then, I thought about information and how it operates and came up with a separate proof that comes to the same conclusion, but through the laws of electrostatics and thermodynamic systems rather than geometry. I took the Nernst equation, one of the basic tools of neuroscience, which calculates the voltage of a neuron, and derived a new equation showing that neurons produce information in a real mathematical and physical way.
And then, I studied a lot more about information and high-dimensional shapes. Theoretical physicists have shown that information always obeys the holographic principle: If there’s a representation of information, then you have information content in higher dimensions from the encoding system. I put all that stuff together and realized that the integration of information encoded in neurons across the brain generates information content in a higher-dimensional space.
Could you sum up what your theory proposes?
The instantaneous integration of electrical activity across a three-dimensional neural network structure, like the human brain, creates information content, which physically exists in a higher-dimensional space. This theory proposes that the mind and the brain operate in accordance with the holographic principle, with the mind forming a rich volume of content while the brain encodes that information representationally.
It probably sounds strange, but that integrated information content has to exist in a higher spatial dimension. That is, a mental state is essentially a hologram of the neural network state. And just like a two-dimensional holographic coding frame generates a three-dimensional image, the brain is a three-dimensional structure that encodes a higher-dimensional projection. I’m arguing that’s why mental states don’t seem to be in the observable universe, though they are generated by and associated with particular patterns of neural activity.
There are a few different ways to reach this conclusion, and I go through four separate routes in the book. The first is the easiest. It’s an intuitive geometrical proof, which considers what spatial dimensions actually are. The logic goes: Spatial dimensions are formed as a function of time. If you have a one-dimensional line changing its position over time, you end up with a two-dimensional area. And if you have a two-dimensional area changing its position over time, you end up with a three-dimensional object. So, any (n)-dimensional shape that is changing over time must produce an (n+1)-dimensional shape.
The brain is a three-dimensional structure, but it’s changing over time, as electrical activity flows through the structure, so it has to create a fourth spatial dimension. That shape contains the integration of all information across that structure over some period of time. Just as we see the shape of a square if we look at the side of a cube — and that surface tells us something about the whole shape — the fourth spatial dimension is tied to the neural network that generated it, but there’s more to it than the surface dimension. I understand that it’s hard to imagine a higher spatial dimension, but it’s worth thinking about that electrical activity flowing through a three-dimensional structure, and what the sum of all that should look like.
The sum total of electrical activity fluxing through the network at a given moment corresponds to the total information content of the brain. Integrated over just a few milliseconds, that information content is a cohesive, streaming, multisensory experience, constantly updated in real time, which can only be accessed by the neural network that produced it. That information content is our perceptual experience, and the sum total of those moments over the course of a lifetime is the self.
So, there’s not only one extra spatial dimension, but two. Perceptual experience is the sum of all information in the brain in a single moment, forming a fourth spatial dimension, and all those snapshots of the neural network state over a lifetime, paired with all the accompanying mental states, are summed together to form yet another spatial dimension, which is the model of the self. Our selves are the information sets encoded by our brains, the sum total of data we have gathered about the world and ourselves over the course of our lives.
So, does neural activity give rise to consciousness? Is it the other way around? Or both?
It’s both. The neural activity that comes in through the senses is data about the world, collected through our senses, that are converted into electrical signals. The processing of those electrical signals in the brain gives rise to the mental experience. But if we look more closely at the electrical signals fluxing through the network, we see neurons are not binary — their behavior is probabilistic. Neurons add up multiple signals, and if they reach a certain voltage threshold, they open these voltage-gated ion channels to permit a flood of charged ions into the cell. So, while neurons do send all-or-nothing signals, this event only happens if there’s coincidental signaling that causes the neuron to reach threshold. And in those moments of uncertainty, the neuron is sensitive to quantum-level noise. So, while most quantum effects disappear at a macro scale, they don’t in neurons and neural networks.
This is what makes the nervous system so unique. This is why consciousness arises from the brain but not from other organs or other objects. While the neurons are adding up all these signals, neurons are sensitive to the exact position and momentum of individual ions, inside or outside of the cell. And because it is a fundamental law of the universe that ions have an uncertain position and momentum, every ion in the vicinity of a neuron is subject to that uncertainty principle. As a result, a neuron’s electrical state is the sum total of the probable positions and momentums of every ion in its vicinity. And to take it a step further, a neural network is the sum total of all probable states of every neuron in the system.
This is a biological description of what I described earlier — a double-integration event. All the possible states of the particles in the system, in relation to every neuron in the system, are described probabilistically, and the sum of all those possible physical states of every particle in the system in relation to every neuron is described as a Schrödinger wave function. That complex number-space is an extra spatial dimension, and that is our perceptual experience. The sum of all possible states of the system is also the very definition of information, and that integrated information, essentially a probability density matrix, generates content holographically. So, there are a couple different ways of saying the same thing there.
But the really interesting thing is how that probability set collapses to allow the neural network and every particle within it to take on a defined state. I use the mathematics developed by Max Born and Richard Feynman to show how this should happen. So, the theory assumes that quantum electrodynamical principles are both true and applicable to biological tissue. The result is that collapse of the probability state should cause physical effects in the system itself — a defined change in the behavior of atoms that essentially implements the probability calculations into the thermodynamic system, causing synchronous electrical activity in neurons across the network. Since synchronous activity does occur in biological neural networks, this framework is consistent with neuroscientific observations.
How would we know if your theory is correct?
If the theory is correct, then we should be able to build a quantum computer based on the same principles as biological neural networks. If the structure and operation of our brains does permit physical information processing as I have described, then we have a model for a working quantum computer. If applying these laws toward engineered systems can provide better predictions for the way such systems operate, we can say it’s a useful theoretical framework.
We can also try to disprove the theory using more parsimonious models. Assuming that neural networks create probability sets in higher-dimensional space can explain a lot, but it’s not the simplest hypothesis. If people can show that neurons act purely deterministically, and that mental states can be completely explained by events in the observable universe, then that work would disprove this theory.
If consciousness can be explained physically, where does that leave free will?
I would say it’s not completely free. This theory accounts for exerting will over the physical world because goals imagined in the mental realm can then be implemented into the neural network, and that can ripple out into effects on the muscles and the body. But it also explains why things that we’re used to doing are a lot easier, because those neural network states are more energetically favorable. In your neural network, essentially all possibilities are open, but in practice, the states that have been occupied in the past are more likely in the present.
What about some sort of ethereal soul — does this theory disprove this idea, or can it be compatible with it?
I would sort of equate the concept of a soul with the self. I think we grow as a function of what we do and what we think, and we do become bigger as we interact with others and live our lives, in a real physical way. It’s not some invisible thing that’s not subject to the laws of our physical universe. I think consciousness is very connected to our bodies. And I think our actions are important, because they affect what is possible in our world and what is more probable to happen in the future.
Is it possible that consciousness could outlast our death?
I actually leave that open in the book. That’s a question of whether information is permanent, and I think we just don’t know enough about the physical laws guiding information to know for sure. I talk about neurological diseases where cells in the brain die off, like in Parkinson’s, and a person may not even be able to connect with information that they used to have. I don’t know if that information goes somewhere else or if it disappears. There are some physicists who say information is indestructible. But in what form? We need to know a lot more about information and entropy and heat to answer that question.