Spacetime connectomics for describing and modelling complex dynamics of human brain states

The human brain is poised to maintain reproducible behaviours while simultaneously allowing for flexible adaptability. Such characteristics are realised in complex systems endowed with metastable dynamics. The central tenet of this Thesis is to investigate rich and spontaneous brain dynamics in terms of their spatial and temporal properties in different brain states, namely the psychedelics induced and depressive state. In Chapter 2, I formalise spatio-temporal dynamics as a temporal trajectory evolving in a space of phase-locking substates. I define temporal characteristics of those substates in terms of fractional occupancy, life times and transition probability. I show functional relevance of the phase-locking substates, recurrence across recording sessions, and a moderate reliability for potential personalised fingerprinting. These results provide theoretical underpinning of spatio-temporal dynamics and motivate use of this particular representation in diagnosis and prognosis of various brain states. Chapter 3 builds on Chapter 2 studying how whole-brain models, fitted to the spatio-temporal description of brain states, can be used for intervention with psychedelics for treatment-resistant depression. I build and validate whole-brain models for responders and non-responders to treatment and by systematic perturbation stimulation I analyse how different regions permit (or not) a transition to healthy brain function. The results demonstrate a subset of regions unlikely to transit to the healthy state. Moreover, these regions correlate with the 5HT2a and 5HT1A density maps known to be mediators of the psychedelic neuropharmacology. This study is a proof-of-concept of the viability in using whole-brain modelling of the brain's spatio-temporal representation, to inform potential avenues in clinical treatments. In Chapter 4, I analyse the psychedelic state induced by N,N-Dimethyltryptamine (DMT) from a perspective of Connectome Harmonic Decomposition. Here, spatio-temporal dynamics are described in terms of temporal contributions of spatial frequencies realised on the structural connectivity of the brain. The results show changes in the harmonic spectrum in line with previous findings with psilocybin and Lysergic Acid Diethylamide (LSD). Furthermore, the DMT-induced state broadens the repertoire of connectome harmonics and enhances signatures of critical dynamics in the brain. These results provide further evidence of whole-brain spatial and temporal changes under the psychedelic-induced state and offer further evidence of the brain being tuned closer to critical dynamics. In Chapter 5, I consider functional substates implicitly derived from the brain's communication structure. I call this method HArmonic DEcomposition of Spacetime (HADES). I demonstrate how large proportion of the data can be reconstructed with just a few of the spatial substates and their temporal contributions. By using the same dataset as in Chapter 4, I show suppression of functional hierarchies and an increase of global integration under the influence of DMT. This method promises an exciting avenue for the analysis of gradient-like and locally specialised functional activity in time. The proposed and applied methods in this thesis demonstrate the necessity of describing brain dynamics across space and time from a complexity science perspective. I show how psychedelics-induced states such as DMT, psilocybin and LSD, change spatial and temporal characteristics of brain activity towards more integrative and flexible functioning, and outline how whole-brain models might inform future clinical interventions for treatment of depression.