Event Abstract Back to Event AN INVESTIGATION OF GLOBAL NETWORK STRUCTURE IN BIPOLAR DISORDER Stefani O'Donoghue1, 2*, Louise Emsell1, 3, Camilla Langan1, Natalie Forde1, 4, Dara M. Cannon1, 2, 5 and Colm McDonald1, 2 1 National University of Ireland, Galway, Psychiatry, Ireland 2 National University of Ireland, Galway, Galway Neuroscience Center, Ireland 3 University Hospitals Leuven, Belgium, KU Leuven & Radiology, Belgium 4 University Medical Centre Groningen, the Netherlands., Psychiatry, Netherlands 5 National University of Ireland, Galway, Anatomy, Ireland Introduction: Network analysis is a novel technique used to identify brain connectivity patterns. Metrics of integration and segregation allow for brain networks to be quantified into neurobiologically meaningful measures. Aim: This study utilises graph theory metrics to investigate properties of brain communication, as well as permutation testing to determine cluster based effects between groups. This analysis aims to elucidate disrupted anatomical wiring as an attribute of psychiatric illness. Methods: Participants were recruited as part of the Galway Bipolar Study (43 Healthy Controls & 42 subjects with Euthymic Bipolar Disorder). Diffusion MRI data was acquired on a 1.5 Tesla Scanner. Connectivity matrices were produced using ExploreDTI. Network metrics were generated using the Brain Connectivity Toolbox in MATLAB. The Network Based Statistic toolbox is a method to identify an experimental effect at the cluster level. FWER corrected p-values are calculated for each component using permutation testing. Results: Properties of integration and segregation revealed between group differences in metrics of Characteristic Path Length [p=0.017], Global Efficiency [p=0.015] and Clustering Coefficient [p=0.047]; with greater brain network disorganization in patients relative to controls. Network analysis was carried out through the NBS with M= 5000 permutations at p< 0.05, revealing a contrast in patients relative to controls across varying (t-statistic) thresholds[(t=2, p=0.015); (t=2.5, p=0.017); (t=3, p=0.020)], in which a single disconnected sub-network comprising multiple dysconnections was identified for each threshold. Conclusion: Metrics that characterize brain communication revealed greater disorganization in patients relative to controls. The NBS findings demonstrated consistency across t-statistic thresholds among occipital and default mode network regions. This study, applying novel network based statistics to MRI and DTI data, indicates that disrupted neuranatomical connectivity is a trait based feature of bipolar disorder. Acknowledgements This research was supported by the Hardiman Research Scholarship. We gratefully acknowledge the participants of this study and the radiographers of the University College Hospital Galway Magnetic Resonance Imaging department for their support during data collection.