# A novel approach to exploring youth non-suicidal self-injury heterogeneity: individual differential psychopathology network analysis

## TL;DR

This study introduces a new method to understand the differences among youth who engage in non-suicidal self-injury by analyzing individual psychopathology networks.

## Contribution

The study introduces the individual differential psychopathology network (IDPN) to identify distinct subgroups of youth with non-suicidal self-injury.

## Key findings

- 881 out of 2,376 students (37.1%) exhibited non-suicidal self-injury.
- Two distinct clusters of NSSI were identified with differing severity and psychopathological features.
- Personality traits, depression, and family environment were strongly linked to NSSI severity.

## Abstract

Non-suicidal self-injury (NSSI) is a common behavioral problem among children and adolescents. Previous studies of NSSI have been mostly group-based and lacked specific characterization of individuals with NSSI.

Using convenience sampling, we surveyed all students from three junior high schools in a county in China, totaling 2,376 participants (mean age 13.66, SD 0.98). Assessments included NSSI, anxiety, depression, personality traits, and family environment. Based on the network template perturbation approach, we employed three steps—constructing the reference network, constructing the perturbed network, and computing the individual differential psychopathology network (IDPN). The IDPN was then constructed from questionnaire scores to capture the degree to which abnormal individuals deviate from the normative level. K-means clustering was then applied to explore the internal heterogeneity of NSSI.

Among 2,376 students, 881 (37.1%) exhibited NSSI. Following IDPN construction, we selected 8 characteristics for clustering analysis based on significant changes in at least 2% of the samples. The elbow method indicated 2 clusters. Fisher discriminant analysis showed a classification accuracy of 95.8%, reflecting a good clustering effect. Severity of NSSI in Group 1 was lower than in Group 2, with scores for 7 out of 8 characteristics also lower in Group 1, except for “Control-Organization.” NSSI was associated with personality traits, depression, and family environment, with stronger connections between individual features linked to higher NSSI severity.

We introduced the concept of IDPN in psychometrics, which can reveal relationships among individual characteristics and identify distinct patient subgroups. Further research is needed to confirm its reproducibility and generalizability.

The online version contains supplementary material available at 10.1186/s12991-025-00606-5.