Validation of the Babbly language development app

This study will validate the accuracy of a new artificial intelligence (AI) powered application, which has been developed to track communication and cognition during the babbling stages of development in the first year of life. Once validated, this app will be an asset to both parents and clinicians. It will be used by parents and will be suitable for informing speech-language pathologists and other clinicians about potential developmental communication/cognitive delays during this period. Parents will upload videos of their infants’ communicative interactions and the app will flag children who show signs of an atypical trajectory in their development. Red flags will facilitate both the early assessment of very young children, and appropriate therapies to be put into place. The validation process will include two stages: (1) teaching the AI to identify typical communication behaviors in infancy; (2) testing whether the AI performs accurately with both typically developing infants and those at-risk for atypical development. These will entail, first, manually coding videos for communication behavior (a time-consuming process), then assessing whether or not the AI can accurately categorize these infant behaviors (a time efficient process). Ultimately, this app will enable the earliest identification of language delays in Canadian children.

Faculty Supervisor:

Monika Molnar

Student:

Glynnis Elizabeth DuBois

Partner:

Babbly

Discipline:

Medicine

Sector:

Information and cultural industries

University:

University of Toronto