Psychosis and Illusory Social Experiences: Phenomenology, Cognitive Mechanisms and Social Functioning

Studies on the global burden of disease (e.g., Murray et al., 2013) show that psychosis ranks among the top medical and mental health conditions causing disability. A recent shift in psychosis research has focused on why delusions and hallucinations are predominantly social and relational, often involving illusory social agents perceived as if engaged in real interactions-typically experienced as voices or persecutors (Bell, 2013; Wilkinson & Bell, 2016; Bell et al., 2017). These agents are now becoming a target for treatment (Craig et al., 2017). Yet, the social phenomenology of delusions remains underexplored, and dominant cognitive models fall short in accounting for these social features. Furthermore, socio-cognitive measures only modestly explain functional outcomes in psychosis.
This thesis aims to bridge that gap by advancing our understanding of the phenomenology and socio-cognitive mechanisms underlying delusional symptoms. It proposes the social agent representation framework as an organizing principle of social cognition. This framework posits that disruptions in the cognitive systems responsible for representing and detecting social agents may underlie the experience of illusory social agents in psychosis.
To explore and expand this framework, the thesis addresses four core questions, each examined in a separate empirical chapter:
Chapter 2: Are social themes common across delusions? Two meta-analyses investigate the prevalence of delusional themes across cultures and contexts. One focuses on established assessment scales, the other includes ad hoc and clinical classifications more common in non-Western settings. Both analyses examine clinical and demographic associations and find that social themes are prevalent across diagnoses, cultures, and time periods, with closer relationships such as family or relatives being more commonly involved as illusory social agents than more distant connections like neighbours, friends, or associates.
Chapter 3: Are illusory social agents present in delusional content, and who are they? This qualitative study analyses medical records from a large patient sample in Central London to identify the prevalence and identity of illusory social agents in delusions, extending the literature beyond persecutory themes. This study showed that the majority of delusions are socially themed, frequently involving individual and groups of explicit social agents, typically humans.
Chapter 4: Is hyperactive agency detection a cognitive mechanism driving social delusions? Using interactive game-theory tasks, this experimental study tests agency detection in people with psychosis. Results suggest those with psychosis and paranoia tend to over-detect human agency, in high social information contexts, pointing to biased agency detection as a key cognitive mechanism in psychosis.
Chapter 5: What is the functional and population-level impact of hyper-agency detection? Using agent-based modelling, which is a computational simulation technique, this research chapter simulates how illusory social agents influence social functioning. The results show that such experiences may reduce social success in cooperative contexts, they may be advantageous in anti-cooperative ones-suggesting that heightened agency detection could be adaptive in certain environments.
The studies in this dissertation present a cohesive body of evidence supporting the central thesis that socio-cognitive dysfunction in social agent representation is central to delusional phenomenology in psychosis. This framework bridges the lived experience of delusions with underlying cognitive processes, namely agency detection, offering a clearer account of their social nature. The findings also support the use of interventions like cognitive remediation, social skills training, and sensory modulation to target these deficits, laying the groundwork for future research into the mechanisms driving delusional thought and behavior.