Consciousness: here, there and everywhere?

Abstract

The science of consciousness has made great strides by focusing on the behavioural and neuronal correlates of experience. However, while such correlates are important for progress to occur, they are not enough if we are to understand even basic facts, for example, why the cerebral cortex gives rise to consciousness but the cerebellum does not, though it has even more neurons and appears to be just as complicated. Moreover, correlates are of little help in many instances where we would like to know if consciousness is present: patients with a few remaining islands of functioning cortex, preterm infants, non-mammalian species and machines that are rapidly outperforming people at driving, recognizing faces and objects, and answering difficult questions. To address these issues, we need not only more data but also a theory of consciousness-one that says what experience is and what type of physical systems can have it. Integrated information theory (IIT) does so by starting from experience itself via five phenomenological axioms: intrinsic existence, composition, information, integration and exclusion. From these it derives five postulates about the properties required of physical mechanisms to support consciousness. The theory provides a principled account of both the quantity and the quality of an individual experience (a quale), and a calculus to evaluate whether or not a particular physical system is conscious and of what. Moreover, IIT can explain a range of clinical and laboratory findings, makes a number of testable predictions and extrapolates to a number of problematic conditions. The theory holds that consciousness is a fundamental property possessed by physical systems having specific causal properties. It predicts that consciousness is graded, is common among biological organisms and can occur in some very simple systems. Conversely, it predicts that feed-forward networks, even complex ones, are not conscious, nor are aggregates such as groups of individuals or heaps of sand. Also, in sharp contrast to widespread functionalist beliefs, IIT implies that digital computers, even if their behaviour were to be functionally equivalent to ours, and even if they were to run faithful simulations of the human brain, would experience next to nothing.

Keywords: awareness; causation; cerebral cortex; existence; mind body problem; neuronal correlates of consciousness.

Figure 1.

Behavioural (BCC) and neuronal correlates of consciousness (NCC). The top row shows a schematic diagram of a binocular rivalry experiment. A horizontal red grating is shown to the left eye and a vertical green grating to the right eye throughout the experiment (courtesy of Naotsugu Tsuchiya and Olivia Carter). The subject does not see a juxtaposition of both stimuli but experiences either the red grating or the green one, switching back and forth every few seconds. Even if the stimuli do not change, what one sees consciously does, as is inferred by the subject's report. The bottom row shows the results of an experiment using magnetoencephalography (MEG), in which the red grating was flashed at one frequency and the green one at another. Yellow indicates areas of the cortex (seen from the top) that had more power at the frequency of the red grating when it was experienced than when it was not. The cyan lines indicate increased coherence (synchronization) between distant brain regions associated with experiencing the grating (from [9]).

Figure 2.

Six instances in which it becomes progressively more difficult to infer the existence of consciousness, since the behavioural repertoire and the underlying mechanisms (brains) differ substantially from that of typical persons able to speak about their experiences (figure 1).

Figure 3.

Axioms and postulates of integrated information theory (IIT). The illustration is a colourized version of Ernst Mach's ‘View from the left eye’ [84]. See also the mechanisms in figure 4.

Figure 4.

A didactic example of how to calculate the quality and quantity of consciousness given a system of elements in a state. On the upper left are three gates with binary states (either ON or OFF: ABC = 100; see also figure 3) that are wired together as shown. An analysis based on the postulates of IIT [80] reveals that the system forms a complex. The complex in its present state specifies a quale—a conceptual structure that is maximally irreducible intrinsically. The quale is presented both as the set of maximally irreducible cause–effect repertoires (concepts) specified by each mechanism (top) and as a two-dimensional projection in which each concept is a ‘star’ in cause–effect space (bottom). Cause–effect space or qualia space is a high-dimensional (here, 2 × 8 dimensions) space in which each axis is a possible past (in blue) and future (in green) state of the complex, and the position along the axis is the probability of that state. Each concept is a star whose position indicates how a mechanism composed of a subset of elements affects the probability of past and future states of the system (its cause–effect repertoire, which specifies what the concept contributes to experience) and whose size (φ^max) measures how irreducible the concept is (how much it contributes to experience). In IIT, Φ^max—a non-negative number—measures the intrinsic irreducibility of the entire quale, how much consciousness there is—the quantity of experience. The ‘form’ or shape of the quale (constellation of stars) is identical to the quality of the experience. Different shapes correspond to different experiences: they feel the way they do—red feeling different from blue or from a headache—because of the distinct shapes of their qualia.

Figure 5.

IIT makes several predictions about which systems can experience anything—how much and in which way—and which systems, even complicated ones, have no experience, remaining ‘in the dark’. IIT implies that consciousness is graded (a); that aggregates are not conscious (a, right panel); that strictly feed-forward systems are not conscious (b, right panel), even if they are functionally equivalent in terms of their input–output operations to feedback networks that are conscious (b, left panel); that even accurate biophysical simulations of the human brain running on digital machines would not be conscious like us, but would be mere aggregates of much simpler systems (transistors and the like) having minimal Φ^max (c). The last row (c) shows, from left to right, a human brain (Allen Institute), the IBM Blue Gene P supercomputer, a columnar model of mouse cortex (Blue Brain Project) and a scanning electron micrographic cross-section of 4 NMOS INTEL transistors in a grid.