Colour preferences among individuals with common mental health disorders: a scoping review.

According to China’s seventh national census, the population aged 10–29 years was approximately 320 million, accounting for about 23% of the national population (). The latest Report on the Development of National Mental Health in China (2019–2020) showed that the detection rate of depression among Chinese adolescents was 24.6%. Adolescent depression refers to depression occurring during adolescence, commonly defined as 12–18 years of age. Previous studies have reported that the prevalence of adolescent depression ranges from 1.8 to 7.8%, and that the 12-month prevalence in middle to late adolescence is similar to that in adults, ranging from 4.0 to 7.5% (; ). Non-suicidal self-injury (NSSI) refers to the intentional and direct destruction of one’s own body tissue without suicidal intent, including socially unacceptable behaviors such as cutting, burning, and hitting oneself. Among hospitalized adolescents, approximately 30–45% engage in NSSI (). In community samples, the prevalence of NSSI among adolescents ranges from 9.9 to 35.6% (; ). NSSI not only increases the risk of suicide but also causes physical harm, including scarring and wound infection (). Beyond psychosocial factors, neuroimaging research in adolescents with bipolar disorder has identified resting-state functional connectivity patterns associated with risk and resilience for self-harm (). Some adolescents with depression report addiction-like experiences related to NSSI, such as recurrent urges, loss of control, repeated engagement despite negative consequences, and temporary relief or pleasure following self-injury (; ). The addiction framework is theoretically supported by negative reinforcement models, in which self-injury rapidly reduces aversive affect and thereby strengthens the behavior over time, and by clinical descriptions of craving, tolerance-like escalation, and withdrawal-like tension before self-injury (; ). Nevertheless, this construct is not a formal independent diagnosis in DSM-5 or ICD-11, and the addiction framework remains debated. In the present study, the term “NSSI addiction-like features” is used to emphasize that the outcome reflects clinically relevant addictive characteristics measured by a validated self-report subscale rather than a formal diagnostic category. The addiction subscale of the Ottawa Self-Injury Inventory (OSI) was selected because it was developed based on substance-dependence criteria and has been used in prior studies to operationalize addictive features of NSSI (; ). Recent research has increasingly applied nomograms and machine-learning approaches to identify adolescents at risk for NSSI. For example, developed a machine-learning model for NSSI risk among adolescents in western China, constructed XGBoost and random-forest models for NSSI among Chinese adolescents with major depressive disorder, and developed a longitudinal nomogram for NSSI in youth with depression. Other nomogram studies have also focused on NSSI behaviors among adolescents with depression or specific student populations (; ). These studies provide important evidence for early identification of NSSI behavior, but most have treated NSSI as a binary behavioral outcome rather than focusing on addiction-like features of NSSI. Therefore, the incremental contribution of the present study is its focus on addiction-like features of NSSI among adolescents and young adults with depression, its use of clinically accessible psychological and family-related variables, and its explicit evaluation of calibration, optimism-corrected performance, multicollinearity, and decision-curve net benefit.