Dynamical complexity and metastability in resting-state neuroimaging

Student thesis: Doctoral ThesisDoctor of Philosophy


When we conceive the brain as a complex system, we acknowledge that the brain is composed of a multitude of components that interact in a nonlinear manner, which gives rise to selforganisation and emergence. Self-organisation implies that there is no overarching control element in the system, but nevertheless, patterns form in space and time reflecting the collective behaviour of the system. Emergence is a phenomenon where a feature of the system is due to the synergistic nature of components and their interactions in the system, and which cannot be derived or explained by the sum of the interactions. Simply put, it is the whole minus the sum.

Metastability is a dynamical phenomenon in brain activity which allows an interplay between regions coordinating their activity to achieve a generalized goal, and retaining their autonomy to perform a specialized function. It is a concept that links stability and flexibility, integration and segregation, and the competitive balance between these complementarities appears to be disrupted in psychiatric disorders.

Schizophrenia is a debilitating disorder which affects around 1% of the population where patients experience positive symptoms such as hallucinations, delusions, and thought disorders, negative symptoms such as avolition and anhedonia, and cognitive symptoms such as deficits in working memory and executive function. Schizophrenia is often considered a disconnection disorder where the ability to integrate information across disparate regions of the brain is impaired. Signatures of metastability in brain activity could plausibly reflect differences in integration between groups of healthy individuals and those with a diagnosis of schizophrenia.

Functional connectivity (FC) in the brain, can be elucidated from resting-state functional Magnetic Resonance Imaging (fMRI). fMRI provides a unique opportunity to observe brain activity in a non-invasive manner. Whilst static FC estimates statistical relationships across brain regions over a complete scanning session, signatures of inherent dynamic phenomena such as metastability cannot be derived with this static approach. Instead, the fMRI data needs to be analysed either within a time-window or instantaneously to estimate time-varying or dynamic functional connectivity (dFC).

This thesis documents the investigation of resting-state fMRI within a framework of complexity science. It includes a conceptual review of the 4 C’s of complexity, namely connectivity, computation, criticality and coherence, and highlights the universality of metastability across these theories. It also describes 2 studies which investigated the reliability of several established dFC metrics across multiple scanning sessions in healthy young adults, and then retaining the only stable metric, metastability, investigated group and individual level differences between healthy controls and groups with a diagnosis of either early psychosis or chronic schizophrenia. Finally, it presents a preliminary review of the past, present, and future of the phenomenon of metastability. The thesis closes with a discussion that brings together the conceptual and empirical findings supporting the premise of metastability as a candidate neuromechanistic marker of schizophrenia pathology.
Date of Award1 Jul 2023
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorFederico Turkheimer (Supervisor) & Ottavia Dipasquale (Supervisor)

Cite this