Multi-centre study of neuroanatomical abnormalities in individuals at Ultra High Risk of Psychosis

Student thesis: Doctoral ThesisDoctor of Philosophy

Abstract

Individuals experiencing prodromal symptoms of psychosis have a very high risk of developing the disorder ranging from 18%-­‐36% within three years of the first clinical presentation. Currently, it is not possible to predict which individuals will subsequently become psychotic only on the basis of their presenting clinical features. This potentially prevents the selective delivery of specialised clinical interventions to those individuals more likely to develop psychosis, which is desirable, both from an ethical point of view and for a more targeted use of available treatments. Neuroimaging may aid prediction as recent neuroimaging studies suggest that there are neuroanatomical differences in people at ultra high risk (UHR) for psychosis relative to healthy control subjects. Furthermore, within UHR cohorts, those who later develop a psychotic disorder (UHR-­‐T, transition) often show more marked structural alteration than those that do not (UHR-­‐NT, non-­‐ transition). However the findings have been inconsistent and this may partly reflect the use of small samples and different analytic methods. The aim of this doctoral project was to assess brain structure in individuals at UHR of psychosis using a larger sample than in previous studies. This was achieved by combining Magnetic Resonance Imaging (MRI) data from four different scanning sites and using a range of different analytic methods including voxel-­‐based morphometry, voxel-­‐based cortical thickness analysis and multivariate machine learning. The use of these methods allowed a comprehensive investigation of neuroanatomical differences in a large cohort and, between UHR-­‐T and UHR-­‐NT cases in terms of i) regional gray matter volume; ii) cortical thickness; and iii) subtle and distributed patterns of gray matter alterations. Findings suggest that there are neuroanatomical abnormalities that precede the emergence of psychosis within a distributed fronto-­‐temporal network. In addition, UHR and healthy controls are distinguishable at the individual level based on information on the gray matter volume, whereas UHR-­‐T and UHR-­‐NT are distinguishable at the individual level using cortical thickness information. Nevertheless, the accuracies reported remain relatively low to be applied in real-­‐world clinical settings. Results from this project contribute to expanding the available knowledge on the UHR population.
Date of Award2015
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorAndrea Mechelli (Supervisor) & Paul Allen (Supervisor)

Cite this

'