Functional Connectivity and Machine Learning for Psychiatric Drug Development

Student thesis: Doctoral ThesisDoctor of Philosophy


The human brain is a complex biological network consisting of spatially separated but functionally integrated regions. Study of functional connectivity gained immense popularity in recent years providing new insight into the mechanisms underlying complex functions and the fundamental organisation of the brain. This has led to the emergence of new techniques for investigating connectivity, such as the application of pattern recognition techniques and the investigation of network dynamics. While highly promising, the application of these new techniques to pharmacological imaging data has not yet been fully explored.
In this thesis we apply pattern recognition techniques to functional connectivity measures obtained for pharmacological imaging data to discriminate patterns of whole brain connectivity. Furthermore, we demonstrate that consideration of functional connectivity dynamics provides additional insight into the effect of pharmacological interventions. Specifically, we explore the effects of the N-methyl-D-aspartate receptor antagonist, ketamine, on the connectivity within the human brain. We argue that the investigation of connectivity is a more appropriate tool for the investigation of this compound due to the highly distributed pattern of effects, as compared to traditional approaches investigating amplitude effects. We demonstrate the applicability of pattern recognition techniques for the discrimination pharmacological states using measures of regional connectivity over the whole brain, using network interactions and through the inspection of network dynamics. We expand upon traditional approaches in our investigation, introducing a new approach to investigate network effects and temporal dynamics of connectivity organisation.
Date of Award2015
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorMitul Mehta (Supervisor), Orla Doyle (Supervisor) & Michael Brammer (Supervisor)

Cite this