Postpartum depression screening through artificial intelligence: preliminary data through the Talking About algorithm


postpartum depression
artificial intelligence
Edinburgh Postnatal Depression Scale
deep learning
Speech Emotion Recognition
Talking About algorithm

How to Cite

Fanos, V., Dessì, A., Deledda, L., Lai, A., Ranzi, P., Avellino, I., Marinaro, F., Petza, S., Pintus, R., Oliverio, G., Vitale, S. G., Angioni, S., & Colangelo, A. (2023). Postpartum depression screening through artificial intelligence: preliminary data through the Talking About algorithm. Journal of Pediatric and Neonatal Individualized Medicine (JPNIM), 12(2), e120222.


Postpartum depression (PD) is the most widespread perinatal psy­chiatric disorder, also representing the most frequent non-obstetric birth-related complication. 

From an epidemiological point of view, it has an average prevalence of 17-18% worldwide. This psychiatric disorder may have long-standing effects on the health of both the mother and the child, but also on the relationship with the partner (including paternal PD). Therefore, an early diagnosis is fundamental to treat this disorder immediately and avoid such complications. 

Talking About, by the company GPI (Trento, Italy), is a project focused on voice analysis as a medium to access human emotions. It consists of a series of Speech Emotion Recognition (SER) algorithms. The aim of the study is to evaluate the application of the artificial intelligence (AI) algorithm Talking About on the mothers’ emotions analysis. Talking About investigates the unconscious aspects of voice that usually cannot be controlled or voluntarily modified, aiming at identifying the subjects’ emotions. Thus, all bias, characterising all classic screening questionnaires, should be neutralised, achieving a sharper overview of the mothers’ emotional state. 

The mother’s emotional results are displayed in 5 main categories: 2 positive, 3 negative. 

This study has engaged a total of 154 mothers who gave birth at the “Policlinico Universitario D. Casula” and/or carried out a pediatric examination at the “ambulatorio SOS MAMI” (which is a PostNatal Care Service). They underwent both the Edinburgh Postnatal Depression Scale (EPDS) test and the Talking About voice test. 

Despite this study’s sample limitations, our preliminary data related to PD symptoms identification are promising and encouraging, leading the way to further investigations related to the application of AI as a PD screening support. Indeed, further studies are needed to improve our knowledge on this topic and possibly apply this tool in clinical practice in the future, even considering PD of the father.