Classroom Cries Exhibit More Temporal Structure than Non-Cries at One Year

• In Event: 1-064 - Emotion in the Voice: Leveraging Automated Assessment to Study Communication and Regulation in Early Childhood

Integrative Statement

Crying in young children is a communicative emotional process. Recent research indicates higher levels of temporal structuring among infant non-speech-related than speech-related vocalizations in familial contexts (Abney, et al., 2017). However, little is known about crying in other contexts. In the United States, almost a third of children under five years of age attend a fulltime childcare center making this a central, but understudied, context for communicative development. We compared early cries and non-cries in an intervention preschool to better understand similarities and differences between these communicative vocalizations in this central developmental context.

We conducted 13 school-day recordings over 10 months among 10 infants in an intervention preschool classroom in which infants were below the 30th percentile in expressive vocabulary. Infants had an average of 6.0 observations (SD=3.9), and a mean age of 12.95 months (SD=3.50) at the start of data collection. Audio from individual children was recorded using Language ENvironment Analysis (LENA) Digital Language Processers (DLPs) worn by each child in a LENA vest. The mean duration of child recordings was 3.71 hours (SD=.68). We quantified each child’s vocalizations as non-cries (CHN/CHF) or cries (CXY) using LENA PRO V3.4.0 software. Prior to conducting paired-sample t-tests (two-tailed unless noted) and correlations, we aggregated over observations to create mean measures for each child.

The mean number of child cries (296.76, SD=94.28) and non-cry utterances (256.64, SD=64.34) did not differ, t(9)=-1.66, p=.13 (Figure 1A). Likewise, the mean duration of cries (.87 seconds, SD=.06) and non-cry utterances (.87 seconds, SD=.03) did not differ, t(9)=-.082, p=.94 (Figure 1B).
Both cry (M=.75, SD=.08) and non-cry utterances (M=.72, SD=.06) were highly bursty, meaning that infants typically clustered cries and non-cries rather than spacing them periodically over the recordings. Confirming Abney et al.’s home-based results, cries were more bursty than non-cries, t(9)=2.077, p=.034, one-tailed; Figure 1C).

We next investigated temporal dependency (Messinger et al., 2017), an index of predictability. Temporal dependency was operationalized as the correlation between the durations of successive cries, and of successive non-cries. The temporal dependency of cries (.15, SD=.06) was higher than the temporal dependency of non-cry vocalizations (.08, SD=.05), t(9)= -2.3, p=.04 Figure 1D. In other words, cry vocalizations exhibited a higher level of temporal structuring (e.g., a longer cry predicts a longer cry) than did non-cries.

Individual cry and non-cry vocalizations did not differ in frequency or duration. However, the sequencing of cries were more temporally structured than that of non-cries. Cries exhibited higher burstiness—a higher tendency to occur in clusters—than non-cries. In addition, cries exhibited higher levels of temporal dependency than non-cries. That is, cry durations were more predictive of subsequent cry durations than non-cry durations were predictive of subsequent non-cry durations. Together, these results suggest that cries are a less flexibly produced vocalization than non-cries.

Objective analysis of large corpora of vocalization data afforded a new look at infant communication in the often-ignored preschool context. Ongoing data collection involving physical tracking of infants will afford developmental analyses, and modeling of the sociality of cries and non-cries.

Authors

• Daniel Messinger, University of Miami
  Presenting Author

• Stephanie Custode, University of Miami
  Non-Presenting Author

• Lynn K Perry, University of Miami
  Non-Presenting Author

• Samantha Mitsven, University of Miami
  Non-Presenting Author

• Mary Anne Ullery, University of Miami
  Non-Presenting Author

• Lynne F Katz, University of Miami
  Non-Presenting Author

• Laura Vitale, University of Miami
  Non-Presenting Author

• Emily Barbara Prince, University of Miami
  Non-Presenting Author

• Adriana Valtierra, University of Miami
  Non-Presenting Author

• Neil Johnson, George Washington University
  Non-Presenting Author