Investigating white matter microstructure predating and following youth cannabis use, using diffusion-weighted magnetic resonance imaging from the IMAGEN longitudinal consortium

Robinson, Emily Anne

Item

Investigating white matter microstructure predating and following youth cannabis use, using diffusion-weighted magnetic resonance imaging from the IMAGEN longitudinal consortium

Robinson, Emily Anne

Author

Robinson, Emily Anne

Abstract

Cannabis is one of the most commonly used psychoactive substances worldwide, with 228 million users, or 4.4% of individuals aged between 15 and 64, in 2022 alone. Adolescence is commonly a period when substance use starts and is also when substances are used at a higher rate than in the general adult population, with 5.5% of young people aged 15-16 also using cannabis in 2022, compared to 4.4% of adults. Meanwhile, the strength and availability of cannabis products have been increasing over the last decade. Of relevance, the main psychoactive compound of cannabis, delta-9-tetrahydrocannabinol (THC), binds to endocannabinoid (eCB) receptors located in the brain to exert its psychoactive effects. Importantly, the cannabinoid receptors are part of a complex endocannabinoid system that plays a key role in regulating adolescent neurodevelopment, including white matter pathways that connect brain regions to ensure efficient neural communication. Therefore, the adolescent brain – including white matter integrity – may be vulnerable to changes in relation to cannabinoid exposure. Emerging evidence has examined the microstructural integrity of white matter fibres using advanced neuroimaging techniques including diffusion-weighted MRI (dMRI). The preliminary dMRI body of work suggests that cannabis use, especially early onset or heavy or frequent use, may have a negative impact on the white matter microstructure of the human brain. However, at the time of the thesis commencement, major limitations of the literature were outstanding and affected the understanding of the association between cannabis use and white matter microstructure. Firstly, an up-to-date systematic synthesis of the dMRI literature to date was lacking, preventing a systematic understanding of the association between cannabis use and white matter microstructure. Additionally, no longitudinal dMRI studies have examined white matter microstructural differences in individuals who use cannabis compared to controls, prior to and following the onset of cannabis use. Study 1 comprises a systematic literature review (SLR), pre-registered in PROSPERO and reported as per PRISMA guidelines. The SLR summarised the dMRI evidence to date on white matter differences in adolescent and adult individuals who use cannabis, compared to controls. It included 30 dMRI studies, encompassing 1,457 individuals who use cannabis and 1,441 controls. Across the reviewed cannabis using samples, the average age of first use of cannabis was 15.4 years, the regular use of cannabis (as defined by each study) started at a mean of 16.1 years, with an average use duration of 8.2 years. There was consistent evidence of group differences in primarily diffusion tensor imaging (DTI) metrics. However, the findings regarding the location, direction and the type of DTI metric were heterogenous. The most commonly reported group differences in white matter integrity were seen in selected tracts - the corpus callosum (CC), and superior longitudinal fasciculus (SLF) - and as a function of distinct DTI metrics - Fractional Anisotropy (FA) and Mean Diffusivity (MD). There was also emerging yet mixed evidence that FA differences correlated with cannabis use parameters (e.g., age of onset, and duration of cannabis use), suggesting dose-related effects of cannabis on white matter. Additionally, the dMRI literature was limited by the lack of longitudinal studies examining brain integrity prior to the onset of cannabis use; a limited assessment of adolescent samples, and other methodological issues. These included: the use of unclear/heterogenous cannabis use metrics, inconsistently matched samples, or use of outdated neuroimaging analysis techniques. As such, a series of two dMRI experiments were conducted in this PhD thesis, to address key limitations of the research to date in adolescent samples. Indeed, as all participants reported being cannabis naïve at baseline, we were able to examine for the first time (to the author’s knowledge) white matter changes before and after cannabis use onset, a key novelty of this thesis. Experiment 1 aimed to investigate for the first time white matter differences (i.e., FA) predating and following adolescent cannabis use onset, as measured by Tract-Based Spatial Statistics (TBSS), controlling for alcohol and nicotine dependence, as well as levels of sensation seeking. The secondary aim was to investigate associations between white matter alterations and with chronicity of cannabis use (e.g., age of onset, frequency), levels of alcohol and nicotine use, and personality variables (e.g., impulsivity) related to substance use. Analyses were run controlling for alcohol and nicotine dependence, as well as levels of sensation seeking. A sample of 55 adolescents was selected from the IMAGEN longitudinal dataset. The sample comprised 27 individuals who had no history of any substance use at age 14 but who commenced cannabis use by age 19 (termed ‘cannabis users’), and 27 controls who had not commenced using cannabis by age 19. These groups were matched on general ability index, pubertal development, impulsivity scores, alcohol, and tobacco use at baseline age 14, and on age at both baseline and follow up. The cannabis group started using cannabis at 16 years of age and were consuming cannabis on average 7.8 days per month. The results from Experiment 1, indicated no significant group differences at baseline before the onset of cannabis use, or in the average white matter across time (i.e., average of baseline and follow up for the cannabis group, vs average of baseline and follow up for the control group). Also, there were no significant differences between how the groups changed over time (i.e., change from baseline to follow up for the cannabis group, vs change from baseline to follow up for the control group). The lack of significant group differences precluded the running of correlations with substance use metrics. These non-significant effects might have reflected a true lack of difference; or dose-response issues, whereby the low-frequency of cannabis use reported by this sample were insufficiently high for white matter differences to be identified. Alternatively, small differences may have been present which could be detected only with more precise measures of white matter integrity, or with larger sample sizes. Experiment 2 was designed to address key limitations of previous dMRI research and the first experiment. It comprised the first investigation of white matter microstructure in adolescents who use cannabis, compared to controls via the novel dMRI analysis tool Fixel Based Analysis (FBA). Indeed, FBA provides a more precise metric of white matter microstructure than traditional DTI tools, especially in areas encapsulating crossing white matter fibres. Using the same sample of in a well-matched cannabis users and controls as Experiment 1, the primary aim of Experiment 2 was to investigate differences in white matter microstructure prospectively using FBA. This was undertaken using both exploratory whole brain analysis and hypothesis driven ROI-analyses focused on the SLF and CC, accounting for alcohol and nicotine dependence, and sensation seeking as covariates. The secondary aim was to investigate correlations between white matter integrity differences and cannabis use metrics, levels of alcohol and nicotine use, and personality variables associated with substance use. The results from Experiment 2 were non-significant when using FBA whole brain analyses. Specifically, there were no significant effects of group across time (i.e., average of baseline and follow up for the cannabis group, vs average of baseline and follow up for the control group), or group-by-time on FBA metrics (i.e., fibre density (FD), fibre cross-section (FC), or fibre density cross-section (FDC)). The results remained non-significant both before and after controlling for covariates. Meanwhile, significant results emerged using region of interest (ROI) analysis. Specifically, there were a small number of voxels implicated in significant group by time effects on the CC and SLF. Specifically, cannabis users showed greater change over time in FC in the CC and the SLF, whereas controls remained relatively stable. However, there were no significant correlations between FC in the SLF and CC, and behavioural variables. Analysis of baseline differences also indicated that there were also no significant whole brain differences between prospective cannabis users and controls, both before and after controlling for covariates. However, ROI analyses indicated that FDC SLF at baseline was significantly higher in prospective cannabis users than controls. Additionally, FD in the left SLF-II was significantly higher in cannabis users and controls at baseline, after controlling for AUDIT, FTND, and sensation seeking, and this was negatively correlated with days of alcohol use at follow-up. Overall, the dMRI literature to date to date suggests that earlier onset and heavy cannabis use by both adolescents and adults is associated with differences in white matter microstructure. However, there is mixed evidence for alterations associated with low-frequency cannabis use, with emerging differences evident in specific white matter fibre tracts (e.g., the SLF and CC). The findings from Experiments 1 and 2 suggest that white matter differences are associated with low-frequency cannabis use during adolescence are not detectable using TBSS (Experiment 1) but were detected using FBA in select white matter fibre tracts (e.g., CC, SLF), but primarily driven by white matter microstructure differences evident prior to onset (Experiment 2). However, the findings from the two experiments are the first (to the author’s knowledge) to measure white matter microstructure prospectively both before and after the onset of adolescent cannabis use. As such, further replication research is needed to draw reliable conclusions regarding predisposing factors for cannabis use onset. Additionally, white matter differences in populations of reporting heavy cannabis, compared to controls, likely differ from those from populations with low levels of cannabis use. As such, future research should focus on dose-response effects, comparing the varying impacts of heavy, moderate and low-level use on white matter microstructure, and neuroanatomical outcomes in general. Overall, more longitudinal multimodal neuroimaging research in young people who use cannabis is needed to identify those adolescents who might be more vulnerable or resilient, to experiencing cannabis use related harms, and the possible long-lasting implications for their mental health and wellbeing.

Keywords

cannabis, white matter microstructure, fixel based analysis, tract based spatial statistics, IMAGEN consortium, neuroimaging, magnetic resonance imaging, diffusion MRI