Repository logo

Complex, yet Resolved: Time-, Frequency-, and Information-resolved Approaches to Brain Function and Human Consciousness



Change log



Different perturbations of the brain’s delicate functioning, ranging from transient pharmacological interventions to severe trauma, can result in altered states of consciousness. To illuminate how the neurobiology and organization of the human brain support consciousness, we need to identify changes in brain function that accompany alterations in conscious state. However, the brain is a paradigmatic example of a complex system, raising the question: which aspects of its complex functioning and architecture should be the focus of our investigation? Traditionally, the quest for the “neural correlates of consciousness” has been framed in terms of spatial localisation: which brain regions are most relevant for consciousness? Complementing this extensive body of work, in my thesis I consider three alternative ways of conceptualising brain function (quantified from functional MRI), and how it may support consciousness. First, I adopt a time-resolved perspective, decomposing brain activity into predominantly integrated or segregated patterns of dynamic functional connectivity. Building on my previous work in anaesthesia and disorders of consciousness, I show how the dynamic interplay of functional integration and segregation is reshaped by the classic serotonergic psychedelic, LSD. Second, I consider a frequency-resolved perspective, decomposing functional brain activity into patterns of structure-function coupling across scales: the harmonic modes of the human connectome. This “connectome harmonic decomposition” of brain activity reveals a generalisable neural signature of loss of consciousness, whether due to anaesthesia or brain injury. A mirror-reverse of this harmonic signature characterises the altered state induced by LSD or ketamine. Connectome harmonics provide a robust indicator of consciousness across datasets, correlating with physiological and subjective variables. On the theoretical side, neuroscientific theories postulate that consciousness depends on the integration of information by a “global workspace” of brain regions. However, these accounts treat “information” as a primitive, whereas the recent framework of information decomposition has shown that Shannon information is actually a composite of several more fundamental kinds of information, including synergistic information, which is available only when a set of sources are considered jointly, and redundant information, which is available from multiple individual sources. Demonstrating the importance of disentangling these different kinds of information, I develop a framework for information-resolved analysis of brain activity, based on information decomposition. Combining functional and diffusion MRI, PET, and transcriptomics, I show that higher cognitive systems in the brain leverage the efficiency of synergistic information, whereas redundant interactions are predominantly associated with modular, structurally-coupled sensorimotor systems. Finally, by explicitly taking into account these fundamental kinds of information, I formalise the “global workspace” architecture in information-theoretic terms, revealing that both anaesthesia and disorders of consciousness induce a breakdown of synergistic integration in the brain’s Default Mode Network. Conceptually, these results contribute to reconciling two prominent theories of consciousness, the Global Neuronal Workspace Theory and Integrated Information Theory. Overall, viewing the brain as a time-, frequency-, and information-resolved complex system offers fruitful new ways to understand the brain’s functional architecture, laying the foundations to map the rich landscape of human consciousness.





Stamatakis, Emmanuel
Menon, David


consciousness, functional MRI, synergy, information decomposition, brain dynamics, complex systems, anaesthesia, disorders of consciousness


Doctor of Philosophy (PhD)

Awarding Institution

University of Cambridge
Gates Cambridge Trust (OPP 1144)