Author ORCID Identifier
0000-0003-4022-9267
Date of Graduation
5-2020
Document Type
Dissertation (PhD)
Program Affiliation
Neuroscience
Degree Name
Doctor of Philosophy (PhD)
Advisor/Committee Chair
Jair Soares, MD, PhD
Committee Member
Benson Mwangi, PhD
Committee Member
Khader Hasan, PhD
Committee Member
Scott Lane, PhD
Committee Member
Zhongming Zhao, PhD
Abstract
Bipolar Disorder (BD) is diagnosed using the Diagnostic and Statistical Manual (DSM) criteria, which relies heavily on symptomatology. This method, however, lends itself to error due to variance in symptom expression and is further complicated during childhood and adolescence- a period marked by major anatomical and behavioral changes. Therefore, in order to institute early and effective interventions, it is imperative that we develop more objective methods of mood disorder characterization and diagnosis. Two proposed solutions for accomplishing this task include 1) the generation of normative development models to assess BD risk and 2) data-driven clustering of patients based on specific neuroanatomical profiles, i.e. biotypes. While the normative development model was unable to quantify BD risk at an individual level, it nevertheless emphasized the heterogeneous nature of both healthy and BD development. The clustering analysis, however, was successful at parsing the variance in the BD sample which resulted in the identification of two distinct BD biotypes. Whereas the BD clusters mapped onto specific anatomical and neurocognitive patterns, symptom-based DSM subtypes were not associated with any empirical measures of mental health. Hence, these findings are testament to the potential of unsupervised and unbiased computational methods in the future of BD diagnostics.
Keywords
bipolar disorder, neuroimaging, machine learning, biotypes, psychiatry