HITLAB and Babbly announce results of external validation study
Stan Kachnowski • September 12, 2023
Thrilled to Share HITLAB’s Latest Breakthrough research testing the efficacy of AI in healthcare for infants. (this post was NOT generated by AI) .
At HITLAB, our mission has always revolved around harnessing innovative technology to improve healthcare outcomes for all 8 billion people -around the world. Many thanks for the ongoing support for our work from the April Smith Hirak of HHS, Stephan Konya from ONC, and Dr. Mohammad Baby from the UN.
I am excited to announce the results of our recent collaboration with Babbly, a company deeply committed to early language development in children.
Early detection of language delays can have a profound impact on a child’s literacy skills and overall development. One critical indicator is missed babbling milestones, which can signal potential developmental issues. Infants who experience delays in babbling during their first year are at risk of having smaller vocabularies and delayed language development later on. Furthermore, a lack of babbling can serve as an early indicator of developmental disorders such as autism spectrum disorder or apraxia. Identifying these deviations early on is crucial for effective intervention and improved outcomes.
Our efficacy verification study with Babbly focused on independently validating their AI algorithm for classifying various infant vocalization patterns. Babbly’s algorithm is designed to detect four key infant vocalizations: cooing, single syllable babbling, canonical or reduplicated babbling, and variegated babbling. To assess its accuracy, we used real-world recordings of infants aged 4-16 months, and compared the algorithm’s predictions with annotations from three trained human observers.
I’m delighted to report that Babbly’s algorithm demonstrated remarkable accuracy, achieving an impressive F-1 score of 0.91. This high level of accuracy was consistent across infants of different ages and genders, highlighting the algorithm’s versatility throughout preverbal development stages.
In summary, this external study showcases the potential of Babbly’s AI algorithm in early detection of developmental language delays in infants. By accurately identifying babbling pattern deviations, we can intervene earlier and provide clinicians with critical information, enhancing their ability to support healthy language development in early childhood.
A special thank you to the Babbly team Maryam Nabavi, Iva Brunec, PhD, Carla Margalef Bentabol. Andrew Crichton, Blair Fast, Kelly Schott, Sanjana Jadhav, Zilun Peng, including advisors Dr. Deryk Beal, Dr. Dina Kulik, and Talia Leszcz for the terrific collaboration in this groundbreaking research!
Thank you to HITLAB team for all your hard work, Kat Marriott and Vandana Yadav.
Read the white paper
