Objective
Self-injurious behaviour represents a significant public health concern, falling under the umbrella of non-suicidal self-injury (NSSI) and suicidal self-injury, categorized by the presence or absence of suicidal intent. When suicide is unsuccessful, this behaviour is designated as a suicide attempt(SA). Research has identified decision-making deficits in patients with NSSI and SA, characterized by a proclivity for immediate reward and a disregard for long-term benefits. However, the paucity of studies that objectively analyse the mechanisms of the decision-making process in both NSSI and SA patients has led to a lack of understanding of the aetiology of these behaviors. Therefore, this paper proposes to compare the similarities and differences in the counterfactual decision process mechanisms of NSSI and SA patients using a wheel of fortune task with a computational modelling approach.
Methods
A total of 140 subjects completed the following psychological assessments: the Beck Depression Inventory (BDI); the State Trait Anxiety Inventory (STAI); the Barratt Impulsivity Scale (BIS-11); the Beck Suicidal Ideation Inventory (SSI); the Beck Hopelessness Scale (BHS) and the Childhood Trauma Questionnaire (CTQ). These were completed after basic demographic information had been recorded, followed by the Wheel of Fortune task. A second-order cluster analysis was then used to categorize the 105 clinical subjects into three groups: a high suicide risk group (SA, n=29), a high self-harm risk group (NSSI, n=28), and a low self-harm risk group (PC, n=48) based on the presence or absence of SA or NSSI, and the most recent week's 1-5 scores on SSI and BDI. 35 individuals were then formed into a healthy control group (HC, n=35). The demographic information was subjected to analysis using ANOVA, and linear mixed effects analyses were performed using the lme4 package in R.
Results & Discussion
SA vs NSSI vs PC were no significant differences in disease duration, STAI, BHS and CTQ, SA group had higher BDI and BIS scores and more suicidal ideation in the last week. SA vs NSSI vs PC vs HC were significant differences in BDI, SAI, TAI and BIS.
In depressed patients with a history of self-injurious behaviour, the value of expected rewards had a lesser influence on goal-directed decision-making than in the control group (p<0.05). These patients also exhibited a greater degree of risk aversion (p<0.05). This finding is closely related to the irrational suicidal behavioral choices exhibited by patients in acute risk states. In other words, patients with a history of self-injurious behaviors have lower expectations of the future and choose irrational behaviors of harming themselves in order to avoid risk. Further comparisons between the high suicide risk and high self-injury risk groups revealed that there was no significant difference between the effects of risk and expected value on the decision-making processes of the two groups. However, the high suicide risk group showed more agentic counterfactual thinking and was more influenced by expected regret in the goal-directed decision-making process (p=0.06).
Conclusions
Self-injurious behaviour has been linked to altered decision-making patterns, while suicide risk has been associated with specific emotional processing biases.

Counterfactual,,decision-making,self injury,Suicide attempt,Non-suicidal self-injury