Applying Information Theory to Explain Consciousness

Caltech Neuroscientist, Christof Koch, Consciousness: Confessions of a romantic reductionist: describing Giulio Tononi’s theory of consciousness which applies Integrated Information Theory to understand and explain consciousness:

Integrated information theory introduces a precise measure capturing the extent of consciousness called Φ. Expressed in bits, Φ quantifies the reduction of uncertainty that occurs in a system, above and beyond the information generated independently by its parts, when that system enters a particular state. (Remember, information is the reduction of uncertainty.) The parts—the modules—of the system account for as much non-integrated, independent information as possible. Thus, if all of the individual chunks of the brain taken in isolation already account for much of the information, little further integration has occurred. Φ measures how much the network, in its current state, is synergistic, the extent to which the system is more than the sum of its parts. Thus, Φ can also be considered to be a measure of the holism of the network. Integrated information theory makes a number of predictions. One of the more counter-intuitive, and therefore powerful, ones is that integrated information arises from causal interactions within the system. When those interactions can’t take place anymore, even though the actual state of the system remains unchanged, Φ shrinks. High Φ networks require both specialization and integration, a hallmark of neural circuitry in the cortical-thalamic complex. Φ denotes the size of the conscious repertoire associated with any network of causally interacting parts. The more integrated and differentiated the system is, the more conscious it is. Synchronous firing of action potentials among neurons is another means of integration.

At the core of our humanity is our consciousness. A word we can't even clearly define or understand but all too familiar with. To be human, is to be conscious. That is, to be aware of our existence, mortality and self in this universe. As such, there have been many attempts by philosophers and scientists to explain the conscious brain.

The computer scientist Doug Hofstadter speculates that at the heart of self-consciousness is a paradoxical and self-referential strange loop, akin to the Escher print of a pair of hands drawing each other. If so—and I’m skeptical, because the mind has difficulty self-referencing or recurring much beyond “I am thinking about myself thinking”—this strange loop is incidental to consciousness per se. Rather, self-consciousness is an evolutionary adaptation of older forms of body and pain consciousness. –Christof Koch, Consciousness: Confessions of a romantic reductionist

And even more exotic theories to explain the unconscious part of the brain that’s in charge of most of our vital bodily functions.

Most of what we do and think and feel is not under our conscious control. The vast jungles of neurons operate their own programs. The conscious you – the “I” that flickers to life when you wake up in the morning – is the smallest bit of what’s transpiring in your brain. Although we are dependent on the functioning of the brain for our inner lives, it runs its own show. Most of its operations are above the security clearance of the conscious mind. The “I” simply has no right of entry. -David Eagleman, Incognito

Perhaps as the paragraph above suggests, attempting to understand consciousness is like trying to explain what exists beyond the borders and horizons of our universe. That said, the attempt described in the first paragraph above is one of the most compelling and fascinating theories I have read by Giulio Tononi (See his full paper here). Tononi is a neuroscientist at University of Wisconsin and his brilliant theory attempts to explain consciousness through the prism of information theory.

To put it in very simple terms, Tononi suggests that for a brain or a system to be conscious, it needs to have specialized regions that are also integrated. To illustrate, imagine a company that has many specialized departments with their own specific and unique projects (Picture Apple Inc with its many special departments: iPad, iPhone, Mac, iOS, AirBook, etc) and focus while at the same time those departments are very well connected to one another and collaborate very effectively. The more specialized and more integrated these elements are, the more conscious a system or brain becomes according to Tononi.

High Φ networks require both specialization and integration, a hallmark of neural circuitry in the cortico-thalamic complex. Φ denotes the size of the conscious repertoire associated with any network of causally interacting parts. The more integrated and differentiated the system is, the more conscious it is. -Consciousness: Confessions of a romantic reductionist

What’s also fascinating about this theory is that it explains the unconscious brain. According to Tononi’s theory, what we refer to as the unconscious brain are those elements that are also very specialized in their functions and responsibilities, however, not integrated or networked with other areas or elements.  For various reasons evolution did not favor the integration of these elements.  For example, the isolation of the element controlling the heart rate or digestion must have favored survival instead of leaving it up to the conscious, integrated part of the brain to constantly adjust these automatic bodily functions. To illustrate this point picture a company like Google or Apple again with their secret labs and projects (Google X) that perform very vital functions but do not necessarily integrate or collaborate with all other elements of their network. The theory goes on to explain how each experience has a very unique perception structure in the brain and hence no experience can truly be repeated or for that matter recalled from memory.  What we call memory of an event is a snapshot of that event and rarely captures every part of the experience.

Integrated information theory not only specifies the amount of consciousness, Φ, associated with each state of a system. It also captures the unique quality of that experience. It does so by considering the set of all informational relationships the considering the set of all informational relationships the underlying physical system is capable of. That is, the way in which integrated information is generated determines not only how much consciousness a system has, but also what kind of consciousness it has. -Consciousness: Confessions of a romantic reductionist

Another interesting implication of this theory is that consciousness is not black and white. All systems even artificial ones pose some degree of consciousness which depends on the specialization of their many elements and the complex integration and networking of them. The more specialized and integrated these individual elements become the more conscious that system can become. This notion could mean that the internet could one day become conscious and self-aware. After all the internet is made up of highly specialized nodes, computers, websites, apps, etc, and it is also highly integrated.

All the computers on the planet’s face that are interconnected via the Internet—a few billion. Each one is built out of hundreds of millions of transistors. The total number of transistors throughout the Web (of the order 1018) is a thousand times bigger than the number of synapses in a single human brain (up to 1015). The typical gate of a transistor in the central processing unit is connected to a mere handful of other gates, whereas a single cortical neuron is linked to tens of thousands of other neurons. In other words, neuronal tissue achieves a degree of information integration that is difficult to mimic in two-dimensional silicon technology. Still, the Web may already be sentient. By what signs shall we recognize its consciousness? Will it start acting on its own in the near future, surprising us in alarming ways by its autonomy? -Consciousness: Confessions of a romantic reductionist

A self-aware and conscious artificial intelligence such as the internet is hard to imagine but if this theory is correct, such an evolution of the internet is inevitable and why stop there. Why not a conscious universe or cosmos?

If this theory and its implications sound a bit crazy, let’s remember our place in the life of our universe. A universe that is most foreign to us much like our own brain, most of it made up of dark matter we can’t even observe. It’s been around for at least 15 billion years only about 200,000 of which we’ve been around for as modern humans or homo sapiens and a very small percentage of which we’ve actually been conscious of our and its existence. This means for billions of year the universe or at least our solar system existed with very low consciousness or no conscious observer as we would define one. Just large bodies trapped in gravitational orbits of other stars, going around for billions of years with no purpose or direction, simply obeying the laws of physics. If the chaos of the Big Bang could give rise to the to the modern day self-aware and conscious human brain then the internet becoming conscious and self-aware should not only be inevitable but also not surprising.