Solving the 'real problem': An approach to demystifying consciousness
By Thomas Sainsbury
Anil Seth, Professor of Cognitive and Computational Neuroscience at the University of Sussex, recently gave a seminar at the Center for Developmental Neurobiology as part of the 2021 Neureka! Seminar series. In this seminar Anil discussed his lab’s theoretical and experimental work, providing insights into the neural underpinnings of consciousness. In particular, he showed that viewing the brain as an inference machine, through predictive processing, can help to explain certain subjective qualia of experience.
Consciousness is defined as the subjective experience that shapes our mental lives. In that sense, consciousness is something that all are deeply familiar with. For example, many of us know what it is like to feel the pain of toothache, see the colour green or experience the melodies of a birdsong. In fact, consciousness defines our existence, and “without it there is no world, no self: there is nothing at all” (Seth, 2021). Despite it being evident that consciousness is an emergent property of the inner workings of the brain, much of neuroscience has tended to avoid the study of consciousness directly. Instead, it has focused on understanding how the brain gives rise perception, cognition, learning and behaviour, without ever referencing consciousness. This is because the study of consciousness is plagued by a question “why does consciousness need to exist for these processes to occur”? As such this question has been dubbed the “hard” problem of consciousness by David Chalmers (Chalmers 1995).
In contrast, Seth argues that we can make progress in understanding various properties of consciousness without ever needing to explain its existence. Seth calls this the “real problem” and argues that it is analogous to the study of other phenomena such as the study of life in biology (Seth, 2021). Here, rather than trying to explain life itself, biologists sought to understand properties of living systems such as metabolism, homeostasis, respiration and reproduction. In doing so, much of the mystery of life has been removed through an understanding that it is an emergent property of these processes acting together in concert. Likewise, Seth’s work aims to reductively map first-person accounts of subjective conscious experience onto observed biological mechanisms. To do so, he divides the study of consciousness into distinct phenomena.
The first of these, conscious level, refers to how aware of conscious experience an individual is. Whilst the level of consciousness is altered in sleep, conscious level should not be confused with wakefulness (Seth, 2021). For instance, our dreams can contain vivid conscious experiences, yet we are not awake. Alternatively, in some pathological states, such as vegetative states, individuals can undergo typical sleep-wake cycles without having conscious awareness. Perhaps the most obvious change in conscious level will be recognisable to anyone who has ever undergone surgery; they will have experienced a sudden loss in their awareness as the general anaesthetic reaches their brain, sending them into an unconscious state. Members of Seth’s lab used this abrupt change to understand what features of neural activity may be related to the level of conscious awareness, by recording cortical activity in the brains of subjects while they were awake and when they were under general anesthesia with EEG (Schartner et al., 2015). This revealed striking differences in the organisation of their neural activity, with anesthetized subjects exhibiting simple oscillatory like patterns propagating all over the cortex. Awake subjects, on the other hand, displayed more complex dynamics, with ripples appearing and disappearing over the cortex at different spatial and temporal scales. This difference in complexity can be quantified using metrics similar to those that compress digital photos into JPEG files. Importantly, this same metric can also distinguish between dreamless and dreamful sleep, indicating that it is not simply a metric of wakefulness. This is significant because one of the first steps in being able to actually understand consciousness as a phenomena relies upon the ability to measure it.
Another major focus of Seth’s work is in understanding what gives rise to subjective experience of perception, which he refers to as the content of consciousness. At first examination, it may be tempting to view these sensations as faithful representations of what is actually there in the external world. However, this is not always the case. Take the optical illusion in Fig 1 as an example, the checkered square labelled B appears to be much lighter than the square labelled A. For Seth, these types of illusions reveal something very interesting about the nature of consciousness and brain. It shows that perception is not simply the product of sensory information coming in. Instead, sensory information is combined with the brain's predictions, through what is known as “predictive processing”. Under this theory the brain acts as an inference machine, generating predictions of the sensory world based on prior experience. As such, conscious perceptions can be thought of as controlled hallucinations, which provide the brain's best guess about what is actually there in the external world.
As an experimental demonstration of a similar phenomena, Seth’s lab has utilized a “binocular rivalry” experiment where different visual stimuli are presented to each eye, to test how expectation can modify conscious content (Pinto et al., 2015; Chang et al., 2015). One eye is presented with a random noise pattern, whereas the other is presented with either an image of a face or house which increases in contrast. The subjects are then asked to report which image they see, and their reaction times measured. Seth’s group found that if subjects are told which image is more likely to appear, their reaction time to report the stimulus correctly is reduced . This indicates that the expectation of the stimulus alters the subjective perception of it. In a similar experiment, members of Seth’s lab found that such expectations were accompanied by an alpha rhythm in the EEG recording of the visual cortex, suggesting that this rhythmic activity may be important for encoding those predictions (Sherman et al., 2016). Currently, Seth’s lab is taking these experiments further by creating virtual reality setups to measure the brain's predictive mechanisms in more detail. In these experiments subjects are presented with virtual objects, which can be manipulated, turned and moved around. Some of these objects obey the normal rules of physics, yet some violate these expectations, such as turning in unpredictable ways (Seth, 2021). The ability to perturb sensory stimuli in this virtual setup will allow Seth’s labs to further examine how prior expectations shape perception and how these expectations can be changed through experience.
Seth and colleagues’ novel approach to consciousness research tackles the “real problem” consciousness instead of the “hard problem” by correlating subjective qualities of consciousness with mechanisms of the brain. In much the same way as biology has demystified the process of life, by describing its subprocesses, Seth hopes that his work can help demystify consciousness.
Chang, A., Kanai, R. and Seth, A., 2015. Cross-modal prediction changes the timing of conscious access during the motion-induced blindness. Consciousness and Cognition, 31, pp.139-147.
Chalmers, D. J. 1995. Facing up to the problem of consciousness. Journal of Consciousness Studies 2: 200-19
Pinto, Y., van Gaal, S., de Lange, F., Lamme, V. and Seth, A., 2015. Expectations accelerate entry of visual stimuli into awareness. Journal of Vision, 15(8), p.13.
Schartner, M., Seth, A., Noirhomme, Q., Boly, M., Bruno, M., Laureys, S. and Barrett, A., 2015. Complexity of Multi-Dimensional Spontaneous EEG Decreases during Propofol Induced General Anaesthesia. PLOS ONE, 10(8), p.e0133532.
Seth, A. 2021 Real problems and beast machines: predictive processing and conscious experience. Neureka! Seminar, KCL
Sherman, M., Kanai, R., Seth, A. and VanRullen, R., 2016. Rhythmic Influence of Top–Down Perceptual Priors in the Phase of Prestimulus Occipital Alpha Oscillations. Journal of Cognitive Neuroscience, 28(9), pp.1318-1330.