Neuroimaging metrics of externalizing disorders: assessing neurobiological features of substance use, withdrawal, substance use disorders, and psychopathy
| dc.contributor.advisor | Shane, Matthew | |
| dc.contributor.author | Denomme, William James | |
| dc.date.accessioned | 2024-06-17T15:41:12Z | |
| dc.date.available | 2024-06-17T15:41:12Z | |
| dc.date.issued | 2024-05-01 | |
| dc.degree.discipline | Forensic Psychology | |
| dc.degree.level | Doctor of Philosophy (PhD) | |
| dc.description.abstract | Neuroimaging research has provided several insights into the neurobiological correlates of externalizing features, notably substance use, substance use disorders, and antisociality. For instance, researchers have paired neuroimaging metrics with machine learning (ML) algorithms to classify externalizing patients from controls and externalizing patients with varying severity and prognoses. In addition, studies have used neural reactivity to drug and food rewards to separate cocaine-dependent participants from non-dependent controls, as well as cocaine-dependent with and without a history of withdrawal symptoms and varying degrees of historical cocaine use and psychopathic traits. However, variability in the classification accuracy of ML models precludes inferences of how well neuroimaging metrics can distinguish externalizing patients and controls. Moreover, variability in the classification accuracy of ML models and the lack of work using modalities outside of cue-reactivity preclude sound inferences on how well neuroimaging can distinguish subgroups of externalizing patients. This dissertation consists of three studies to address these factors. In Study 1, a meta-analysis of 49 ML models with neuroimaging predictors demonstrated that neuroimaging metrics could distinguish externalizing patients and controls with an accuracy of ~80%. Study 1 also demonstrated a classification accuracy of ~79% when distinguishing externalizing patients with severity and prognosis differences. However, it is important to note that most studies included in this meta-analysis validated their results using cross-validation, which may have inflated their classification accuracy. Next, Studies 2 and 3 demonstrated that cocaine-dependent participants were distinct from non-dependent controls in terms of gray matter concentration (GMC) or functional connectivity (FNC) in response to drug and food rewards within the orbitofrontal cortex, middle temporal gyrus, dorsolateral prefrontal cortex, middle frontal gyrus (MFG), and anterior cingulate cortex. These studies also found that GMC and FNC within several corticolimbic regions, notably the MFG, separated cocaine-dependent participants with and without a history of withdrawal and varying degrees of lifetime history of cocaine use and psychopathic traits. These results provide preliminary evidence that neuroimaging metrics could distinguish externalizing patients from control, and separate externalizing patients that are subgrouped by symptomology, severity, and prognosis. The presented work has substantial implications for developing novel assessment protocols and optimal treatment strategies. | |
| dc.description.sponsorship | University of Ontario Institute of Technology | |
| dc.identifier.uri | https://ontariotechu.scholaris.ca/handle/10155/1775 | |
| dc.language.iso | en | |
| dc.subject.other | Neuroimaging | |
| dc.subject.other | Machine-learning | |
| dc.subject.other | Meta-analysis | |
| dc.subject.other | Psychopathy | |
| dc.subject.other | Substance use | |
| dc.title | Neuroimaging metrics of externalizing disorders: assessing neurobiological features of substance use, withdrawal, substance use disorders, and psychopathy | |
| dc.type | Dissertation |