Perich T.,University of New South Wales |
Perich T.,Black Dog Institute |
Perich T.,University of Western Sydney |
Hadzi-Pavlovic D.,University of New South Wales |
And 14 more authors.
Acta Psychiatrica Scandinavica | Year: 2016
Objective: To investigate for subtypes of bipolar depression using latent class analysis (LCA). Method: Participants were recruited through a bipolar disorder (BD) clinic. LCA was undertaken using: (i) symptoms reported on the SCID-IV for the most severe lifetime depressive episode; (ii) lifetime illness features such as age at first depressive and hypo/manic episodes; and (iii) family history of BD and unipolar depression. To explore the validity of any demonstrated ‘classes’, clinical, demographic and treatment correlates were investigated. Results: A total of 243 BD subjects (170 with BD-I and 73 with BD-II) were included. For the combined sample, we found two robust LCA solutions, with two and three classes respectively. There were no consistent solutions when the BD-I and BD-II samples were considered separately. Subjects in class 2 of the three-class solution (characterised by anxiety, insomnia, reduced appetite/weight loss, irritability, psychomotor retardation, suicidal ideation, guilt, worthlessness and evening worsening) were significantly more likely to be in receipt of government financial support, suggesting a particularly malign pattern of symptoms. Conclusion: Our study suggests the existence of two or three distinct classes of bipolar depression and a strong association with functional outcome. © 2016 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Zhan L.,University of Southern California |
Nie Z.,Arizona State University |
Ye J.,Arizona State University |
Wang Y.,Arizona State University |
And 8 more authors.
Mathematics and Visualization | Year: 2014
To classify each stage for a progressing disease such as Alzheimer’s disease is a key issue for the disease prevention and treatment. In this study, we derived structural brain networks from diffusion-weighted MRI using whole-brain tractography since there is growing interest in relating connectivity measures to clinical, cognitive, and genetic data. Relatively little work has usedmachine learning to make inferences about variations in brain networks in the progression of the Alzheimer’s disease. Here we developed a framework to utilize generalized low rank approximations of matrices (GLRAM) and modified linear discrimination analysis for unsupervised feature learning and classification of connectivity matrices. We apply the methods to brain networks derived from DWI scans of 41 people with Alzheimer’s disease, 73 people with EMCI, 38 people with LMCI, 47 elderly healthy controls and 221 young healthy controls. Our results show that this new framework can significantly improve classification accuracy when combining multiple datasets; this suggests the value of using data beyond the classification task at hand to model variations in brain connectivity. © Springer International Publishing Switzerland 2014.