RWE, HEOR & Evidence Synthesis

Anhedonia-Related Wording in Social Media: An Application of Natural Language Processing

Axtria collected posts from various mental health subreddits on Reddit to identify users' mental disorders, focusing on anhedonia, depression, anxiety, bipolar disorder, and borderline personality disorder. Anhedonia, the loss of pleasure in previously enjoyable activities, is challenging to diagnose using traditional methods. To improve diagnosis, Axtria used natural language processing (NLP) techniques to develop a proof-of-concept for classifying anhedonia separately from other depression types. They downloaded all posts from the Reddit anhedonia forum between 2017 and 2022 and prepared the data using tokenization, stop-word removal, and lemmatization. Six unsupervised machine learning algorithms, including sentiment analyses and clustering algorithms, were applied to identify patterns in 9,887 posts, which were then interpreted by human analysts.

This report is a poster presentation of Axtria from ISPOR 2024.

Contact us at connect@axtria.com with any questions.

Complete the brief form to download the white paper

Recommended insights

 An Application of Natural Language Processing

Report

The Agentic AI Revolution in Pharmaceutical Operations
 An Application of Natural Language Processing

Report

Axtria Ignite 2025 Fireside Chat: Charting the Future of Commercialization in an Agentic Era
 An Application of Natural Language Processing

Report

Insights from Axtria Ignite 2025 - Executive Insights: Building a Scalable, Purpose-Driven AI Ecosystem Across the Life Sciences Value Chain