Using Infants' cry Vocalizations to Assess Post-Partum Depression in their Mothers.

• In Event: Poster Session 06
  In Poster Session: PS 06 Topic - Biological Processes: Psychophysiology

Abstract

Postpartum Depression (PPD), a patology that affects up to 15% of mothers in high-income countries, reduces attention to the needs of newborns and is among the first causes of infanticide. PPD is usually identified using self-reported measures and therefore it is affected by a social desirability bias, woth mothers unwilling to report it. Previous works have identified the presence of significant differences in the acoustical properties of the cries of infants of healthy and depressed mothers, suggesting that mothers’ behavior can induce changes in infants’ behavior. In this work, cry episodes of infants (N = 56, 157.4 days ± 8.5, 62% firstborn) of non-depressed (N = 27) and depressed (N = 29) mothers (mean age = 31.1 years ± 3.9) have been investigated to assess the possibility acoustical properties of infants’ vocalizations can be identified by a machine learning models as beloging to children of clinically depressed mothers. Acoustic features (fundamental frequency, first four formants, and intensity) are first extracted from recordings of crying infants, then cloud-based artificial intelligence models are employed to identify post partum depression versus non-depression from estimated features. Trained model shows that commonly adopted acoustical features can be successfully employed to identify depressed mothers with good degree of accuracy (89.5%).

Authors

• Giulio Gabrieli, Istituto Italiano di Tecnologia (IIT)
  Presenting Author

• Marc H Bornstein, Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD/NIH)
  Non-Presenting Author

• Nanmathi Manian, Westat
  Non-Presenting Author

• Gianluca Esposito, University of Trento
  Non-Presenting Author