Dissociation, interoception, physical and mental pain and their associations with suicide-related thoughts and behaviors: A series of network analyses and preliminary cluster analysis

Introduction and objectives: Suicide is the fourth cause of death among 15-29 year-olds and university students have been highlighted as a particularly vulnerable group. Other authors have focused on dissociation, interoception and both physical and mental pain as possible suicide-risk factors. Each dimension has already been singularly examined, but from what we know, no study has yet explored all such dimensions jointly, in their interactions and in association with different suicide-related events (SREs). We hypothesized that, at least for a subgroup of people, higher dissociation, interoceptive deficits and altered pain processing can play a role in the transitioning from suicidal thoughts to behavior. The aim of this work is therefore to see if dissociation, interoceptive awareness, physical and mental pain perception interact with each other and possess specific associations with suicidal idea- tion (SI), suicidal planning (SP), non-suicidal self-injury (NSSI) and suicide attempt (SA). Materials and methods: A preliminary sample of 507 university students completed an online survey with measures of dissociation (DES-II), interoception (MAIA-2), pain vigilance/awareness (PVAQ), mental pain (Psychache Scale), physical and psychological pain (Visual analogues), alexithymia (TAS) and fearlessness about death (ACSS-FAD), as well as a self-report adaptation of the C-SSRS for the screening of SI, SP, NSSI and SA history. A first network-analysis was performed with all the variables and the SREs. As the SREs were highly inteconnected, the detail of their relationships was also examined in a separate network. Then to understand how each SRE was associated with the other dimensions, a network for each SRE with all the other variables was performed. A preliminary cluster-analysis with random forest machine learning algorithm was also attempted to check for the existence of a dissociative cluster among the SI group. Results: Participants were aged 18-29 years and 31.7% (161) of them reported lifetime SI, 10.9% (106) reported SP, 35.5% (180) reported NSSI and 9.9% (50) SA history. Inside the SREs-only network, SI was a better predictor of SA than SP. In network with the other variables, but without the other SREs, SI was connected with MP, DES and FAD, but MP was the best predictor. Conversely, SA was also associated with both MP and DES, but DES was the best predictor. Among the group with SI, 5 clusters were extracted, among which 3 clusters of suicide attempters emerged. Conclusions: The specific sample does not allow to generalize results to other populations. Our hypotheses would benefit from control for other relevant dimensions and longitudinal data. However, complex relationships among SREs, dissociation, pain and interoception have emerged, the role of MP and DES has proven pivotal, while a possible dissociative cluster requires further data.