Sackler Centre for Consciousness Science

Theory and Modelling

Understanding consciousness requires new theory as well as new experiments.

By bridging disciplines from mathematics to psychiatry, the Sackler Centre is uniquely well placed to advance theory and experiment together.  Sackler Centre researchers have been actively contributing to the development of theoretical consciousness science in a number of important ways.

A key concept driving our theoretical approach is that of explanatory correlates of consciousness (ECCs). This idea builds on the currently-dominant approach of looking for ‘neural correlates of consciousness’, which are brain regions or processes that correlate with conscious level (e.g., the difference between being awake or asleep) and conscious content (a particular conscious experience).   But correlations are not themselves explanations, nor do they isolate causal relationships between brain activity and consciousness.  An ECC, by contrast, is a brain process that causally accounts for specific aspects of conscious phenomenology, thereby going well beyond correlations to provide real explanations of how brain, mind, and consciousness relate.

A good example of an ECC is that of ‘neural complexity’.  This is based on the insight that conscious experiences are both differentiated (each is one among a vast repertoire of possibilities) and integrated (each conscious scene is unified).  This suggests the underlying neural processes should also be simultaneously differentiated and integrated – i.e., ‘neurally complex’ in mathematically specific ways.  We are developing new approaches for measuring and modelling this kind of complexity, based on concepts of Lempel-Ziv complexity, Shannon entropy, synergy, causal density and integrated information. Datasets analysed include EEG recordings from subjects undergoing general anaesthesia, and intracranial depth electrode recordings from awake and asleep epileptic patients. We are also seeing whether these theoretically-driven approaches will help us better understand what underlies the pathological loss of consciousness in coma and the vegetative state.

Another prominent line of work has to do with 'predictive processing' or the 'Bayesian brain'.  On this view, perception is a process of inference on the (hidden) causes of sensory signals. At the Sackler Centre, as well as exploring the empirical implications of these ideas, we are generating new theory in several different directions.  We have extended predictive processing to interoception (the sense of the body from within), to account for emotion and subjective feeling states.  And we have developed versions of 'counterfactual' predictive processing in which the brain makes predictions about possible actions, and we use these ideas to help explain aspects of conscious phenomenology like 'objecthood' and 'presence'. We also work on the mathematical basis of predictive coding, with reference to the 'free energy principle'.

The key to linking brain activity to consciousness – or to any brain function – is to be able to decipher causal interactions among different brain regions from neuroimaging data. One useful tool for doing this is Granger causality, a measure of ‘directed functional connectivity’ which is based on precedence and predictability. Put simply, one signal A is said to ‘Granger cause’ a different signal B if (and only if) A contains information that helps predict the future of B, over and above information already in the past of B. At the Sackler Centre we have been pioneering the theory and application of Granger causality to data from neuroscience. We have written the standard analysis software in the field, which is freely available as a fully-documented MATLAB toolbox, and we are examining how the method can applied in common neuroimaging contexts like functional MRI (fMRI), EEG, and intracranially recorded brain signals. 

References:

Barnett, L.C., and Seth, A.K. (2014). The MVGC multivariate Granger causality analysis toolbox: A new approach to Granger causal inference. Journal of Neuroscience Methods 223:50-68.

Barrett, A.B., and Seth, A.K. (2011). Practical measures of integrated information for time series data. PLoS Computational Biology, 7(1):e1001052

Barrett, A.B. (2014). An integration of integrated information theory with fundamental physics. Front. Psychol. 5(63).

Schartner, M., Seth, A.K., Noirhomme, Q., Boly, M., Bruno, M.-A., Laureys, S., and Barrett, A.B. (2015) Complexity of multi-dimensional spontaneous EEG decreases during propofol induced anaesthesia. PLoS One 10(8):e0133532

Seth, A.K., Barrett, A.B., and Barnett, L.C. (2015). Granger causality analysis in neuroscience and neuroimaging. Journal of Neuroscience 35(8):3293-3297

Seth, A.K. (2014). A predictive processing theory of sensorimotor contingencies: Explaining the puzzle of perceptual presence and its absence in synaesthesia. Cognitive Neuroscience (target article) 5(2):97-118

Seth, A.K., Barrett, A.B., and Barnett. L. (2011). Causal density and integrated information as measures of conscious level. Phil Trans R. Soc. A. 369:3748-3767