STOP project: detection of suicidal ideation on social media
STOP is a research project that studies mental health issues on social media through Artificial Intelligence to find patterns related to the high risk of suicide or depression. A recently published study in the Journal of Medical Internet Research shows very promising results. Dr. Ana Freire (principal investigator, Universitat Pompeu Fabra, UPF) and her team, among which are our CVC researchers Dr. Jordi González and Diego Velázquez, were able to detect suicidal patterns with an 85% of accuracy when analysing texts, images and activities on Twitter with Artificial Intelligence.
More than 3,000 persons commit suicide every year in Spain according to the Spanish National Institute (INE). Therefore, suicide can be considered an important public health issue. Engineers, Psychologists and Psychiatrists work together to change the current suicide numbers by studying mental health issues on social media through Artificial Intelligence to find patterns related to the high risk of suicide or depression.
Suicide risk assessment usually involves an interaction between doctors and patients. However, a significant number of people with mental disorders receive no treatment for their condition due to the limited access to mental health care facilities; the reduced availability of clinicians; the lack of awareness; stigma, neglect, and discrimination surrounding mental disorders. In contrast, internet access and social media usage have increased significantly, providing experts and patients with a means of communication that may contribute to the development of methods to detect mental health issues among social media users.
The objective of this project is to apply Artificial Intelligence on social media data to discover suicide-related patterns: suicide ideation, depression, eating disorders. We explore behavioral, relational, and multimodal data extracted from social media and develop machine learning models to detect risk patters. All our data is anonymized to protect the privacy of the users.
Researchers from the Computer Vision Center (CVC), Pompeu Fabra University (UPF), Autonomous University of Barcelona (UAM) and Parc Taulí Hospital, are working on STOP project (Suicide Prevention in Social Platforms) and launched targeted campaigns on social media addressed to users that match a risk profile identified by artificial intelligence. These campaigns offer a 24h hotline for emotional support. Thanks to this project, “Teléfono de la Esperanza” (suicide help phone) calls have increased by 60% coming from social media.
The last results published in the Journal of Medical Internet Research were promising as they demonstrated they can detect suicidal patterns with an 85% accuracy when analysing texts, images and activities on Twitter with Artificial Intelligence.
“Analysing publications on Twitter, in a completely anonymous way, we found that could exist a correlation between the content of the images shared in this social media with the mental health of the users who published it”, explained Dr. Jordi González, CVC researcher.
STOP in the media:
Ayuda de la Inteligencia Artificial para detectar conductas suicidas en la red – El Correo Gallego
IA podría identificar comportamiento suicida – Parada Visual
La inteligencia artificial para detectar patrones suicidas en redes sociales – Entorno Inteligente
Reference: Ramírez-Cifuentes D, Freire A, Baeza-Yates R, Puntí J, Medina-Bravo P, Velazquez DA, Gonfaus JM, Gonzàlez J. Detection of Suicidal Ideation on Social Media: Multimodal, Relational, and Behavioral Analysis, J Med Internet Res 2020;22(7):e17758 doi: 10.2196/17758 PMID: 32673256
- Diana Ramírez-Cifuentes, Ana Freire et al. “Detection of suicidal ideation on social media: multimodal, relational, and behavioral analysis.” Journal of medical internet research 22.7 (2020): e17758. JCR: 5.03. Q1.
- Ríssola, Esteban, Diana Ramírez-Cifuentes, Ana Freire, and Fabio Crestani. “Suicide risk assessment on social media: USI-UPF at the CLPsych 2019 shared task.” In Proceedings of the Sixth Workshop on Computational Linguistics and Clinical Psychology, pp. 167-171. 2019.
- Diana Ramírez-Cifuentes, Marc Mayans and Ana Freire. Early Risk Detection of Anorexia in Social Media. In Proc. of INSCI 2018 – International Conference on Internet Science. 2018. SJR 0.29. Q2.
- Diana Ramírez-Cifuentes and Ana Freire. UPF’s Participation at the CLEF eRisk 2018: Early Risk Prediction on the Internet. Conference and Labs of the Evaluation Forum. CLEF 2018.
- Ana Freire, Joaquim Puntí-Vidal, Montserrat Pàmias-Massana and Michele Trevisiol. Suicide 2.0: Tracking Online Symptoms to Detect Suicidal Behaviour. Salut Mental 4.0. Les noves tecnologies aplicades a la recerca i tractament en Salut Mental. Quart Seminari de la CORE en Salut Mental. 2017.
Website of the project (More information and complete list of publications)