Child Talk: Measuring the growth in communication for individuals with and without language delays

• In Event: Nonverbal communication

Abstract

To create a more meaningful inclusive classroom, it is important to further investigate the engagement across peers in the environment. Researchers reported that certain locations in an early childhood classroom yield more talk that other classroom locations by using two wearable sensors to collect real-time speech and location information (Irvin et al., 2017). Using digital sensor technologies offer improvements in naturalistic movement measurement and analysis. This can also impact clinical decisions and such decisions can be made by practitioners in inclusive settings to support the development of friendships and to facilitate more engaging interactions in various activity locations in the classroom (Altman et al., 2020). It is known that developing Automatic Speech Recognition (ASR) systems for children is far more challenging than for adults (Gerosa et al., 2007), primarily due to various developing (e.g., articulation/pronunciation, physiology/motor skills, vocabulary, and grammar) and environmental (e.g., overlapping speech, ambient noise) factors. Various technology has been utilized thus far to capture child vocalizations, however, there is a lack of literature to support child talk in preschool settings.
Using wearable sensors and supported by speech/language processing technology, this study explores the dimensions of child talk and development over three timepoints for children with and without language delays. Students were equipped with wearable sensors to measure talk and location. This provided the opportunity to explore where and what speech is occurring in various classroom locations. For early childhood, speech sound development often follows a typical pattern and can be observed throughout early childhood development. This study specifically focuses on four children both with and without language delays in an inclusive classroom. Data was collected over three periods of time over the course of 3 months in various natural routines (e.g., art/snack, block, books, dramatic play) in the classroom. Initial results indicated the effectiveness of the ASR and can be viewed in Table 1. Using these initial results, this analysis expands initial findings to determine different rates of growth between children with language delays compared to their peers with typical development. It will also include analysis of how rates of talk vary amongst routines, activities, and/or time of day. Implications for future research include the importance of various inputs that both teachers and children have with one another and the impacts of location measurement of speech development. These findings will lead to a greater impact for understand how speech and language processing technology can support young children’s development and practitioner effectiveness in early childhood settings.

Authors

• Rebecca Hacker, University of Kansas
  Presenting Author

• Satwik Dutta

• Dwight Irvin, University of Kansas

• Jay Buzhardt, University of Kansas

• John Hansen, University of Texas at Dallas