Connectome-based prediction of substance use disorder severity
| Author(s) | Miglin, Rickie | |
| Date Accessioned | 2024-10-29T16:15:55Z | |
| Date Available | 2024-10-29T16:15:55Z | |
| Publication Date | 2024 | |
| SWORD Update | 2024-10-13T19:03:19Z | |
| Abstract | Background: The United States is facing an epidemic of substance use-related deaths, and a substantial number of these deaths are associated with the use of more than one type of drug. There are a variety of biopsychosocial vulnerability factors that can contribute to an individual’s risk for developing and maintaining a substance use disorder (SUD), including social determinants, neurobiological, cognitive, and psychological factors. In order to design more effective intervention strategies for SUD, it will be necessary to create models testing the relative contribution of these risk factors in maintaining SUDs. ☐ Aims: The overarching objective of the present study was to construct an integrated biopsychosocial model of SUD vulnerability, encompassing functional brain connectivity measured with functional magnetic resonance imaging (fMRI) alongside well-established social and psychological risk factors with the aim of assessing their respective contributions towards explaining variability in SUD severity in a diverse sample of community adults. The first aim was to create a model of SUD vulnerability spanning established personality, cognitive, and social determinant factors to examine the relative contribution of these domains in explaining variance in SUD severity. The second aim was to use connectome-based predictive modeling (CPM) to explore functional neural connections associated with SUD severity, and to extract summary statistics representative of neurobiological factors associated with SUD severity. The third aim was to test whether the CPM summary feature statistics remained significantly related to SUD severity above and beyond the variance accounted for by the variables of the previously tested SUD vulnerability model. ☐ Methods: (Study 1) Models were created for Disordered Personality, Cognitive Ability, and Social Determinant latent variables representing the domains expected to contribute to SUD severity. Structural equation modeling was then performed to model the associations of SUD severity with the latent variables (Disordered Personality, Cognitive Ability, and Social Determinant) and Delay Discounting. (Study 2) CPM was employed in a “Discovery” sample of individuals to identify functional connections measured with fMRI predictive of SUD severity and to extract summary statistics. The model identified in the Discovery sample was then applied, unchanged to an External Validation sample in order to test the generalizability of the model. (Study 3) Another SEM was created to test whether CPM-summary statistics remained significantly associated with SUD severity above and beyond the variables measured in Study 1. ☐ Results: (Study 1) The Social Determinants latent variable was the most strongly, positively associated with SUD severity, followed by Disordered Personality and Delay Discounting. SUD severity was not significantly associated with the Cognitive Ability latent variable in the SEM. (Study 2) CPM identified neural network connections predictive of SUD severity in unseen individuals within the Discovery sample, but the model did not generalize to the External Validation sample. (Study 3) The CPM summary statistics did not explain additional variance in SUD severity in the integrated SEM, above and beyond the variance accounted for by other explanatory variables. ☐ Implications: The study’s findings provide support for biopsychosocial models of SUD and highlight the contribution of social determinant risk factors in maintaining SUD severity. Between- and within-network connectivity of the Posterior-Multimodal, Cingulo-Opercular, Default, and Visual networks may also be important in maintaining SUD severity. Additional research is needed to create comprehensive biopsychosocial models of SUDs in order to develop more effective assessments and treatment approaches. | |
| Advisor | Sadeh, Naomi | |
| Degree | Ph.D. | |
| Department | University of Delaware, Department of Psychological and Brain Sciences | |
| DOI | https://doi.org/10.58088/vkfa-fp27 | |
| Unique Identifier | 1483242884 | |
| URL | https://udspace.udel.edu/handle/19716/35402 | |
| Language | en | |
| Publisher | University of Delaware | |
| URI | https://www.proquest.com/pqdtlocal1006271/dissertations-theses/connectome-based-prediction-substance-use/docview/3116068230/sem-2?accountid=10457 | |
| Keywords | Addiction | |
| Keywords | Biopsychosocial model | |
| Keywords | Drug use | |
| Keywords | Substance use disorder | |
| Title | Connectome-based prediction of substance use disorder severity | |
| Type | Thesis |