Functional brain networks in schizophrenia: mapping connectivity and topology at early and late psychotic illness stages

© 2017 Dr. Eleni GanellaSchizophrenia is a severe mental disorder that is characterised by symptoms including hallucinations, delusions and disorganized thought. The cause of schizophrenia remains unknown; however, it is thought that a combination of genetics, environment and altered neurobiology play a role in the emergence and perpetuation of the disorder. Accumulating evidence suggests that disrupted brain network connectivity may in part underlie the pathophysiology of psychosis, and that network connectivity is to some extent genetically determined and heritable. However, there is still much to be learned surrounding the nature of network abnormalities and how they differ in early versus late psychosis. Exploring the underlying neurobiology at discrete clinical stages of psychotic illness creates a framework to evaluate the biological factors that may be contributing to the progression from early psychosis, to more advanced chronic stages of the disorder. This thesis used resting-state functional magnetic resonance imaging (fMRI) to characterise network functional connectivity and topology in early and late psychosis, as well as in a group of unaffected family members (UFM) of individuals with schizophrenia. Resting-state fMRI is a well validated and sensitive tool for probing the intrinsic functional integrity of the brain. Specifically, this thesis used a data-driven approach to map the temporal coherence of fMRI time series (functional connectivity) across the whole brain. To complement the resting-state functional connectivity (rs-FC) analysis, this thesis used graph theory to explore functional network topology. Network topology describes that brains ability to maintain a balance between local processing speed and global integration of information. These methodological approaches were used to investigate network abnormalities in three groups relative to healthy controls; a first-episode psychosis (FEP) group, a treatment-resistant schizophrenia (TRS) group and a group of UFM. This thesis aimed to investigate 1) whether rs-FC and network topology was abnormal in the early FEP stage of schizophrenia relative to healthy controls at two time-points (baseline and at 12-months follow-up); 2) whether rs-FC and network topology was impaired in a chronic TRS group relative to healthy controls; 3) whether abnormal rs-FC and network topology was evident in a group of UFM, and whether any network measure could be characterised as a marker of risk or resilience to psychosis in UFM. Firstly, results showed no evidence of abnormal rs-FC or topology in FEP individuals relative to healthy controls at baseline, or at the 12-months follow-up. Further, longitudinal changes in network properties over a 12-month period did not significantly differ between FEP individuals and healthy controls. Secondly, this thesis found widespread reductions in rs-FC in the TRS group that predominantly involved temporal, occipital and frontal brain regions. The TRS group also showed reduced global network efficiency and increased local efficiency relative to controls. Thirdly, TRS and UFM shared frontal and occipital rs-FC deficits, representing a ‘risk’ endophenotype. Additional reductions in frontal and temporal rs-FC appeared to be associated with risk that precipitates psychosis in vulnerable individuals, or may be due to other illness-related effects, such as medication. Functional brain networks were more topologically resilient in UFM compared to TRS, which may protect UFM from psychosis onset despite familial liability. Together, the body of work presented in this thesis provides a number of novel and unique findings that serve to advance the current state of knowledge regarding the pathophysiology and heritability of psychosis. Specifically, the work demonstrated that the latest most severe stage of psychosis, TRS, is associated with widespread reduced rs-FC, and that milder, yet similar patterns of dysconnectivity were observed in UFM, implying a genetic root to some, but not all of the observed network abnormalities. Network topology differed relative to healthy controls in both UFM and TRS patients, suggesting that functional network architecture is also disturbed in late psychosis, and again, results suggest a genetic/shared environmental basis for this characteristic. Our finding of no significant difference in rs-FC or network topology in our FEP sample suggests that there is a differentiation between biological processes occurring in early and late psychosis with a subgroup of individuals’ rs-FC potentially being unaffected in the FEP stage