Imaging the Relationship Between Cannabinoids, Reward Processing, and Psychosis

Student thesis: Doctoral ThesisDoctor of Philosophy



There are two distinct temporal phases that encompass reward processing: An anticipatory phase to predict an upcoming reward (or loss) in response to a stimulus and a receipt phase (or outcome phase) where the previously predicted outcome was received or omitted. The anticipation of a reward, followed by its receipt, is associated with a prediction error (PE). PEs denote the mismatch between the observed and the predicted outcome value (e.g., an outcome that was better than what was expected). Impairments in reward processing have been described in a number of psychiatric disorders including psychosis. The unifying observation between impaired reward processing and psychosis are alterations in brain dopamine signalling. Therefore, it has been proposed that psychotic symptoms may arise from aberrant reward processing leading to increased spontaneous dopaminergic firing in mesolimbic reward pathways. This creates ‘neurobiological noise’ which are misinterpreted as meaningful leading to the onset of hallucinations and delusions.

Cannabis sativa is composed of over 140 different phytocannabinoids. The most investigated cannabinoids include delta-9-tetrahydrocannabinol (THC) and cannabidiol (CBD). THC has been associated with the induction of transient psychotomimetic effects in healthy and vulnerable people, as well as those with schizophrenia. Moreover, an association between THC content of recreational cannabis with a greater risk of onset and relapse of psychotic disorders has been reported. However, there has also been growing interest in the therapeutic potential of THC for a number of conditions. Pre-clinical and human evidence have suggested that THC may modulate dopamine transmission leading to the onset of acute psychotomimetic symptoms building upon evidence that hyperdopaminergia represents the final common pathway to psychosis. Moreover, there is growing interest in the antipsychotic potential of CBD following evidence that it may oppose the psychotomimetic and neurophysiological effects of THC in healthy individuals complemented by preliminary evidence of antipsychotic efficacy in patients with psychosis in some clinical trials. However, the neurocognitive mechanisms underlying the effects of THC and CBD on the human brain in the context of reward processing remain unclear.


The principal aims of this project were to investigate whether the psychotomimetic effects of THC and the antipsychotic potential of CBD may be underpinned by their effects on the reward system.


1. This thesis first examined the spatial acute effect of THC on human on regional brain activation or blood flow (together termed as activation signal) in a ‘core’ network of brain regions from 372 participants, tested using a within-subject repeated measures design under experimental conditions. It was also investigated whether the neuromodulatory effects of THC are related to the local expression of the cannabinoid-type-1 (CB1R) and type-2 (CB2R) receptors, as obtained from an independent data set of participants obtained from the Allen Human Brain Atlas. Finally, this thesis investigated the dose-response relationship between THC and key brain substrates.

2. Next, A within-subject, crossover, double-blind, placebo-controlled trial across 16 healthy controls was conducted. Administration of a single dose of THC was compared with placebo while participants performed the instrumental probabilistic learning task, an fMRI paradigm. Signed and unsigned prediction errors were estimated by contrasting rewarding stimuli against a neutral control condition.


3. Through meta-analysis of neuroimaging studies, this thesis examined the acute effects of CBD, relative to placebo, on human brain function using SPECT and fMRI while performing diverse cognitive tasks. Subsequently, this thesis examined the relationship between the spatially distributed pooled effects of CBD on brain signal and the distribution of candidate mechanistic targets for the effects of CBD including fatty acid amide hydrolase (FAAH), dopamine D2, serotonin and cannabinoid-type-1 receptors as indexed by their gene expression data (obtained from the Allen Human Brain Atlas).

4. Next, a within-subject, crossover, double-blind, placebo-controlled investigation of 19 healthy controls and 15 participants with early psychosis was conducted. Administration of a single dose of CBD was compared with placebo in psychosis participants while performing the monetary incentive delay task, an fMRI paradigm. Anticipation of reward and loss were used to contrast motivationally salient stimuli against a neutral control condition. Reward receipt (feedback) of monetary rewards was also contrasted against a neutral control.

5. Finally, a randomised placebo-controlled double-blind, parallel-arm trial was conducted across 33 participants at clinical high risk for psychosis. The effect of a single dose of CBD, relative to placebo, on insular cortex brain activity and task performance during the monetary incentive delay task was used to predict changes in therapeutic outcomes to three weeks CBD administration, compared to placebo using linear mixed effects regression modelling techniques.


The first meta-analytic study investigating the effects of THC, relative to placebo, found that THC augmented the activation signal, relative to placebo, in the anterior cingulate, superior frontal cortices, middle temporal and middle and inferior occipital gyri, striatum, amygdala, thalamus, and cerebellum crus II and attenuated activation signal in the middle temporal gyrus (spatially distinct from the cluster with THC-induced increase in activation signal), superior temporal gyrus, angular gyrus, precuneus, cuneus, inferior parietal lobule, and the cerebellum lobule IV/V. Using post-mortem gene expression data from an independent cohort from the Allen Human Brain atlas, a direct relationship was found between the magnitude of THC-induced brain signal change, indexed using pooled effect-size estimates, and CB1R gene expression, a proxy measure of CB1R protein distribution, but not CB2R expression. A dose-response relationship was observed with THC dose in regions including the anterior cingulate/paracingulate, thalamus, and the supplementary motor area.

In the second study it was found that a single dose of THC administration, compared with placebo, attenuated brain activity within the striatum and midbrain while processing signed positive and negative prediction errors. No significant effect of THC, relative to placebo, was observed in the salience network ROI during unsigned PE processing. Moreover, no THC effect compared to placebo was observed on task performance or learning.

The CBD meta-analysis identified the functional modulation of several brain regions, including the medial frontoparietal, midcingulo-insular, pericentral, lateral frontoparietal, and dorsal frontoparietal networks as well as the striatum and cerebellum following CBD relative to placebo. There was a significant inverse relationship between the magnitude of pooled CBD effect on brain activation and expression of FAAH but not the other targets (dopamine D2, serotonin and cannabinoid-type-1 receptors).

The next study that took place during the anticipation phase of the monetary incentive delay task (MIDT) observed no group differences in brain activation between early psychosis patients compared with healthy controls. Attenuation of insula activation was observed following CBD, compared to placebo, in early psychosis. Sensitivity analyses controlling for current cannabis use history did not affect the main results. During the feedback phase, no significant differences in activation were observed between early psychosis with healthy controls or between psychosis-CBD and psychosis-placebo during reward delivery. Psychosis had lower monetary reward and performance accuracy scores, with faster reaction times, and more frequent premature starts, compared to healthy controls. CBD administration was associated with increased performance accuracy relative to placebo.

Finally, within participants at clinical-high-risk for psychosis, a significant inverse interactive relationship was observed between the effect of CBD on left insula brain signal change with positive psychotic symptoms which was not present in the placebo group. A significant interaction was observed between acute drug effect on mean monetary reward in predicting positive psychotic change, however, there was no significant effect of CBD following individual correlation analysis.


The chapters that comprise this thesis suggests that THC and CBD, when compared to placebo, modulate brain signal during task-independent cognitive processing across an array of brain networks. An opposite effect pattern, of modulated brain signal, emerges when comparing the results of the THC (vs placebo) with the CBD (vs placebo) analyses. The opposite effect pattern may be explained, in part, by the second key finding of this thesis which indirectly demonstrated a relationship between the spatial modulatory effects of THC and CBD on brain activation and their distinct molecular targets, CB1R and FAAH respectively. It is likely that the opposite effects of THC and CBD on brain activation is mediated by their relationship to their key, yet distinct, molecular targets.

This thesis also identified that THC and CBD administration modulated the activation of brain regions involved in reward processing which may be related to their respective psychotomimetic and antipsychotic effects. THC, relative to placebo, was associated with attenuated striatal-midbrain activation, during PE processing, while healthy participants performed an instrumental learning task. When combined with current evidence, these studies suggest that alterations in dopamine signalling, induced by THC, can alter PE processing, thereby, inducing psychotic-like symptoms. Meanwhile, CBD, relative to placebo, was associated with attenuated insular activation during the anticipation phase of the MIDT in participants with early psychosis. This thesis also demonstrated that the degree of attenuated insula activation, following acute CBD administration, compared to placebo, may be indicative of better positive clinical outcomes following sustained CBD administration in the clinical high risk for psychosis population. These findings are significant as evidence suggests that atypical insula engagement is a feature of psychotic disorders. Moreover, dysfunction in reward processing is associated with one of the prevailing theories of psychosis that suggests that the attribution of motivational significance or salience to contextually irrelevant or neutral stimuli may be associated with the onset of psychotic symptoms. This is consistent with the idea that psychotic symptoms may be induced by elevated dopaminergic release within mesolimbic reward pathways causing aberrant stimulus-reinforcement. Therefore, a normalising effect of CBD on aberrant insular hyperactivity may have relevance in its proposed antipsychotic effects. Finally, this thesis suggests that antipsychotic drug administration may introduce confounding effects of dopamine antagonism which may partly ameliorate brain activation abnormalities in patients with psychosis, such that they may no longer be detectable in comparison with healthy controls. Therefore, this raises the question as to whether further research is warranted to investigate the effects of CBD on modulating brain activation during reward processing in medicated psychosis participants.

Future studies should probe the relationship between induction positive psychotic symptoms and brain modulation following THC compared to placebo during reward PE processing using larger sample sizes. This increase in power will permit correlational analysis between the modulatory and symptomatic effects of THC, relative to placebo. Such studies should also adopt a multimodal approach, incorporating PET, to investigate any relationship with CB1R binding. This is essential to understand the relationship between CB1R binding, brain activation modulation, and the induction of transient psychotic symptoms following THC. Similarly, future studies investigating the antipsychotic effects of CBD should also adopt a multimodal neuroimaging approach using radiolabelled FAAH in psychosis. This would be invaluable in understanding the mechanism of the antipsychotic effect of CBD. Finally, future studies should investigate neural biomarkers that may predict treatment response outcomes to antipsychotic drug administration using reward-based paradigms. Such studies should also integrate longitudinal information to evaluate the long-term accuracy of the predictions such as transition rates to chronic schizophrenia. Together, these studies will help advance current understanding of the mechanisms that underlie the psychotomimetic effects of THC and putative antipsychotic effects of CBD.
Date of Award1 Apr 2023
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorSagnik Bhattacharyya (Supervisor) & Kelly Diederen (Supervisor)

Cite this