Ephemeral, ineffable, the ghost in the machine, since the days when we could begin to think about our thoughts, we have attempted to describe the experience of consciousness. We have to. Describing and agreeing on this sensation with one another is the only way we can reassure each other that we feel it too, that we aren’t the only ones experiencing sentience. As Descartes described in his “I think, therefore I am” thought experiment, our introspection proves only our own existence, and thus there is no way for us to prove that anyone or anything else experiences this phenomenal characteristic of life. But that may not always be the case. Neuroscience is racing, literally in some instances, to understand consciousness. Although we have not reached an answer yet, several popular theories have emerged over time that may help us to shed light on our existence.
The highest on the list, pun intended, comes the collection of theories that rely on a top-down view of consciousness. Here, mental states are experienced as “conscious” when they are the target of specific higher-order so-called ‘meta-representations’. Your perception of a landscape, for example, is only felt consciously because lower-order signals in your visual cortex are targeted by certain higher-order meta-representations in your anterior cortex, especially the prefrontal cortex. ‘Higher-order regions’ refers here to brain areas dedicated to cognitive processes that are considered to be more advanced, such as analysis and evaluation of signals. The lower-order regions meanwhile are less complex and are usually involved in direct transfer, processing and reaction to incoming information.
These Higher Order Theories are supported by a mixture of lesion studies and functional network studies that link the anterior cortex to metacognition. Furthermore, these theories explain why some signals are experienced consciously, and some aren’t: they can either be targeted by higher order meta-representations or simply cannot be. Take our landscape scene again; your visual cortex network that processes the view of the trees, hills and beautiful sky are targeted by higher-order functional networks that provide the meta-representation of the scene (making you experience it consciously), but the sound of the gate creaking behind you is not experienced consciously because it cannot be targeted at that time.
First proposed by Daniel Bor, the GWT provides a very functional approach to consciousness. Here, a mental state is experienced as conscious when it enters a distributed network referred to as the Global Workspace. How does this happen? By subconsciously shifting our attention to signals, we raise them into the Global Workspace such that they enter the conscious mind. Where higher order theories rely on specific functional regions in the frontal cortex acting on incoming signals, the GWT is by definition globally distributed across the brain, thus promoting cognitive processes from working memory to perception. This theoretical “workspace” depends on both prefrontal and parietal cortices in the brain becoming active and communicating via ultra-fast brainwaves, resulting in a widespread network interconnectivity.
This theory is considered highly functional however, because although it explains effectively how processes become consciously experienced, it does not account for the phenomenal differences of experience between types of cognitive process. For example, GWT explains why a view of the sunset is felt consciously, but does not explain why the conscious feeling of seeing a tree differs from a sunset, or even why no two sunsets are ever seen the same way.
Mathematically derived by Giulio Tononi, IIT attempts to understand consciousness as a measurable phenomenon with a unit of “phi” Φ that is expressed, like a computer, in bits. Phi interprets the amount of information generated as a whole entity in comparison to the parts of the system alone. Consequently, consciousness is a fundamental property of any system as defined by the complexity and state of the system. In the brain, this implies that consciousness arises more from the neuron-dense posterior areas, where by virtue of their network complexity, the value of phi should be greater. However, this theory would also imply that any system, if complex enough, could and already does experience a degree of consciousness, no matter how natural or synthetic. This implies a degree of panpsychism (the belief that everything experiences a degree of consciousness) where even a sufficiently complex n by n grid of XOR gates would produce a phi value greater than that of the human brain. Given that mycelial networks span tens of square kilometres and are deeply complex, where does that leave us on the consciousness sale?
Based on the discoveries of the models that our brains create to represent things in the world from our bodies to physical space, the attention schema describes an internal model generated by the brain to represent the brain. Here, attention is once more key to the theory, and this simplified description of the brain exists to control our attention. Without this attention model we would be overwhelmed by the sheer volume of raw information that enters the brain through various signalling pathways. Consciousness, as a simplified, descriptive model of the brain ensures that our attention is focused, thus ensuring that the information is used purposefully.
This theory aims to answer how consciousness would arise in the first instance; we evolved it to streamline our wandering attention and have control over our actions. Indeed, Michael Graziano, who proposed the Attention Schema theory, enforces this point by highlighting that the process of natural selection by which evolution occurs often generates the most efficient answer to a pressure. By nature of its simplicity (a model of the brain that controls our attention), the attention schema theory of consciousness is thus the most compelling explanation for the appearance of consciousness, he argues. How such a model of the brain could be quantified remains to be seen.
Once more favouring a top-down model of consciousness are the predictive processing theories. Much like GWTs, predictive processing theories rely on studies that have demonstrated damage to prefrontal areas leads to changes in the subject’s conscious experience. Although not a theory of consciousness in its own right, PPTs are based on a well established notion of top-down predictions from higher-order cortical areas and bottom-up prediction errors arising due to signals from lower-order regions. How this works is relatively straightforward, our brains are always making predictions about its environment, these predictions are then compared against the reality of the environment. The predictions arise from so-called higher-order regions, hence they proceed from the top of the cognitive hierarchy downwards. This comparison against the reality of the environment is called prediction error. If the prediction error is low, then our prediction was correct, if it is high, then our prediction was incorrect and the information is updated for better predictions in the future.
This relationship is similar to that of Bayesian inference, and it’s the established nature of the prediction error relationship combined with the mathematically familiar Bayesian probability that makes PPTs a compelling theory of consciousness for many. How this arises as our conscious experience has to do with our brain’s best predictions about the causes of its incoming signals. Our experience of feeling sad is the brain’s best guess based on bottom up signalling on the status of the body and our surroundings, as a result we experience sadness.
It is no doubt evident to many that there is a great deal of overlap between these theories, and this is not entirely surprising as much like religions, some of these are off-shoots and branches from one another. Some are built from a more pragmatic and mathematical framework, whereas others are theories that incorporate more elements of the human experience. It is easy to get lost in the empirical in neuroscience and taking a step back to observe how we perceive our conscious experience before trying to tie it down into one unifying theory is just as important as running the numbers. The attention schema, for example, captures the experience of feeling like the “ghost in the machine”, but this does not make you a separate entity from your body and brain, as this attention can be a part of other theories that incorporate the role of attention, namely the Global Workspace or Predictive Processing. Your beliefs in whether consciousness is unique to our species play a role too. Higher-order theories assume that consciousness is solely the quality of creatures possessing specific higher-order brain regions, while Integrated Information Theories indicate that anything with sufficient complexity may be considered conscious. In his recent book, Dr Anil Seth makes the point that measurement is often a great turning point in scientific fields, and one day perhaps we will be able to measure consciousness in all living things. Maybe then we can finally prove Descartes wrong.
Author: Thomas von Rein