What is consciousness? Monash University is at the forefront of an extraordinary push to unlock the biggest mystery of the human brain to understand consciousness in humans. What is it? How does the brain generate it? When does it begin and end?
The project has expanded internationally with input from a growing team of collaborating researchers across the fields of physics, engineering, robotics, neuroscience and philosophy.
The consciousness project could also potentially unlock the way forward for machine consciousness. The hope is that by building algorithms for consciousness, the machines will be better and we’ll know more about ourselves.
The possibilities are endless. Here, the four key Monash thought leaders driving the project talk us through their work.
Associate Professor Naotsugu ‘Nao’ Tsuchiya, Monash Institute of Cognitive and Clinical Neurosciences
Professor Jakob Hohwy, School of Philosophical, Historical and International Studies
Professor Tom Drummond, School of Electrical and Computer Systems Engineering
Dr Kavan Modi, School of Physics and Astronomy
Why are you studying consciousness?
Nao Tsuchiya: Consciousness is one of the three mysteries in the world: what is the origin of the universe, what is the origin of life, and then what is the origin of consciousness?
My interest has always been how a brain, a physical substance, can generate subjective phenomena or a conscious experience. Though this question is a difficult one, the answer is directly relevant to our lives. What will happen to my own mental life if I get brain damage that prevents me from moving at all, yet I am fully conscious? This is a state called ‘locked-in syndrome’.
And also, now the emergence of highly intelligent robots and artificial intelligence [AI] – if they turned out conscious, can we switch off their consciousness? If they break laws, who will take responsibility? These are relevant and important questions to us as a society.
Tom Drummond: It really is the last truly great problem, and it’s going to take an interdisciplinary approach to solve it. Nao is a neuroscientist, Jakob is a philosopher, I’m an engineer and computer scientist, and Kavan is a physicist. We all have our own way of thinking about machine intelligence and machine consciousness, and human intelligence and human consciousness and how that works, and inevitably that brings great strength, because we’re looking from different angles. But there’s a challenge, because we all speak different academic languages.
Jakob Hohwy: Consciousness is what makes us what we are; it harbours our thoughts, feelings, experiences and memories. It’s astounding that science doesn’t yet know how the brain gives rise to and shapes the conscious experience that fills our every waking moment. I want to know this, and I’m convinced that our highly interdisciplinary approach is the right way forward.
Kavan Modi: Take a black box. You don’t know if there’s a computer or a person inside it. When do you decide the black box is conscious and begin making ethical decisions about how you treat the black box and how you behave with it? This isn’t a theoretical problem; this is something we’ll face soon. It’s a moral question, but it requires a technical solution. Where do we draw the line?
What is consciousness as far as we know?
Nao Tsuchiya: It’s defined in two ways. First in the medical setting, where you have levels of consciousness from none when you’re dead to very high when you’re fully awake. Then there’s the contents of consciousness, which I’m more interested in. That is conscious perception or whatever you feel – vision, hearing, touch, thoughts, senses, all the things you feel. How do some neurons in the brain produce visual sensations while other parts of the brain produce audition?
This question becomes even more difficult to answer once we realise there are massive amounts of activity in our brains which have nothing to with what we’re experiencing at the time – that’s called non-conscious processing.
It’s clear that not all of what a brain processes is relevant in generating what you’re experiencing. Even sensory inputs don’t directly decide what you experience, as is clear from dreaming and mind wandering. These experiences are generated in the brain and aren’t directly connected to sensory inputs.
Jakob Hohwy: The basic unit we need to solve is, ‘What is one experience?’. So, you wake up in the morning and you see a red rose. And you’re conscious of it. So, what’s that? What is it to be conscious and be conscious of something?
And then there’s a whole host of other related problems that are equally difficult around that about the subconscious, about the stream of consciousness, about the unity of consciousness; about altered states of consciousness, about the self. All these problems come crashing down on you immediately when you start thinking about consciousness. But the basic problem is, ‘What is conscious experience? Why are you conscious rather than not conscious?’ That is the starting point.
Tom Drummond: One of the key ideas in consciousness is attention – you become conscious of something when you pay attention to it. For example, my feet are in contact with the floor, and I’m aware of that now that I think about it.
Attention does a couple of important things. It turns a big, multi-dimensional sensory field into a one-dimensional sequence of components that simplifies the problem of recognising and understand what is happening.
So, if I look at a big, complex painting with lots of characters in it doing different things, I can attend to different parts of it in sequence. This turns the painting into a narrative representing simpler things one at a time.
From this I could generate an essay about what’s happening in the painting. So, attention takes a rich sensory field, and by focusing on one part at a time, it turns it into an internal narrative stream, which is one of the key properties of consciousness.
Kavan Modi: At some point, there’ll be a machine that will fool humans. How many humans does it have to fool for us to consider it to be conscious? Understanding consciousness will take us beyond the machine-human separation.
And then in patients with little or lower brain activity. How do you decide on their fate in terms of consciousness. Are they alive or dead?
What does each of you bring to the table?
Tom Drummond: Nao's focus is on how neurons can implement consciousness, the hardware we have in our brains, how attention and consciousness and a lot of the properties of consciousness maps onto neural hardware.
We can learn something about human behaviour and decision-making, and it can work alongside the neuroscience to understand how all this behaviour is embedded in the neurons.
Jakob is interested in theories of the mind, which can explain experience and how that can relate to mathematical theories of how human learning develops. I want to know how we can make this real. How do we make an artificial object with consciousness? If done well, we can learn something about human behaviour and decision-making, and it can work alongside the neuroscience to understand how all this behaviour is embedded in the neurons.
Kavan Modi: There are a lot of crackpot theories of quantum consciousness, and I don't want to be associated with that. I specialise in quantum mechanics and causation, but causation is the thing I’m doing here.
I work on a whiteboard. I don’t play with machines. I know very little about consciousness and cognition, but I know a lot about causal modelling, stochastic modelling [estimating probability with random variations] and building mathematical frameworks to solve problems that seem unsolvable. Where it might go, I have no idea.
It would be cool if we could get some minimalistic models and have fundamental statements – conclusive statements about what we can say about consciousness.
What you often want to do is find something simple. You want to be able to say, 'Within these structures you cannot do this, or this always exists'. You want to be able to put a neat box around things. This is the physicist’s approach.
Nao Tsuchiya: I’m interested in empirically testing a relatively new neuroscience and mathematical theory called integrated informational theory. We want to find the causal explanation between what’s going on in the brain and what we consciously experience at each moment.
With careful experiments and detailed theories, constructed through this collaboration with engineers and physicists, I believe we can make massive progress in this direction.
Jakob Hohwy: Philosophy is the oldest discipline involved in the study of consciousness. Together with my colleague Tim Bayne – an expert on measuring consciousness who’s also part of this project – we’ve worked on the question of whether there are levels of consciousness and what the right scientific method would be for the study of consciousness. I also work on theories in theoretical neuroscience that underlie consciousness.
What’s Monash University's role in international consciousness research?
Jakob Hohwy: There’s critical mass now at Monash for doing this, and a joint track record as well. We’ve worked now for a long time with the neuroscientists, we’ve co-supervised students, we’ve published together. And I have my lab here where I do actual experiments, even though I’m sitting in a philosophy department.
I think Monash is good at this, this openness to collaboration – people bring resources to the table and ideas, then you can go for it and see what happens. So, Monash becomes this centre node for an academic research network of people who are interested in this question.
Monash gave us an internal seed grant to get this international network up and running. Next year we’ll have some activities where we try and get the multidisciplinary, international team together and try and do some science.
We’ve invited a postdoc or senior PhD student from each of the four international labs we’re engaging with to come to Monash for an extended period of time at the same time, and put them in close proximity to each other and to us, and then break new ground in the science of consciousness.
What does the future hold, then, in human-like machines, given what you all know?
Tom Drummond: We’re going to be using AI systems to assist us in all sorts of ways, so we’ll need something closely resembling machine consciousness in order to have a conversation with the machine about why certain behaviours by the machine aren’t acceptable. They need to exhibit interactive, considered behaviour.
Jakob Hohwy: A lot of machine learning is really good at coming up with solutions in very restricted, orderly environments. That’s great for a lot of purposes – industry robots that work within a cage, and that kind of thing.
But what’s characteristic of us is that we work really well in volatile, changing environments, so there’s this imperative to look not just for artificial intelligence but for general intelligence, which is a system that’s able to deal with change, not just in the colour of the screws that it’s looking at, but that is able to understand if the screws have been swapped with apples.
All stuff that we can do really, really easily – and that seems tied to us being conscious, attentive creatures.
To receive a fortnightly email wrap up of stories from Lens.