Detecting Early and Implicit Suicidal Ideation via Longitudinal and Information Environment Signals on Social Media

2025-10-16

AI Summary

Implicit suicidal ideation (SI) on social media is often expressed indirectly through everyday posts and social interactions, posing significant challenges for early detection. To address this, we propose the first multi-source predictive framework that jointly models users’ long-term posting histories and their temporally aligned interactions with socially proximate neighbors. Our method introduces a composite network centrality–based neighbor selection strategy to identify critical social nodes and integrates temporally aligned user–neighbor interaction sequences with a fine-tuned DeBERTa-v3 model. Evaluated on a dataset of 1,000 Reddit users, our approach achieves a 15% improvement in early implicit SI detection over text-only baselines, significantly enhancing risk surveillance capability.

Technology Category

• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Social Networks and Social Media: Social mining and social search on the Web
• User Modeling, Personalization and Recommendation: Social recommender systems and personalization for the social Web
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

On social media, many individuals experiencing suicidal ideation (SI) do not disclose their distress explicitly. Instead, signs may surface indirectly through everyday posts or peer interactions. Detecting such implicit signals early is critical but remains challenging. We frame early and implicit SI as a forward-looking prediction task and develop a computational framework that models a user's information environment, consisting of both their longitudinal posting histories as well as the discourse of their socially proximal peers. We adopted a composite network centrality measure to identify top neighbors of a user, and temporally aligned the user's and neighbors'interactions -- integrating the multi-layered signals in a fine-tuned DeBERTa-v3 model. In a Reddit study of 1,000 (500 Case and 500 Control) users, our approach improves early and implicit SI detection by 15% over individual-only baselines. These findings highlight that peer interactions offer valuable predictive signals and carry broader implications for designing early detection systems that capture indirect as well as masked expressions of risk in online environments.

Problem

• Detecting implicit suicidal ideation from social media posts
• Modeling user information environment and peer interactions
• Improving early risk detection through longitudinal signal analysis

Innovation

• Uses longitudinal user and peer interaction data
• Integrates multi-layered signals in fine-tuned DeBERTa-v3
• Employs composite network centrality for neighbor identification