Neonatal imitation of caregivers at home: Pre-registered analyses

• In Event: Poster Session 13
  In Poster Session: PS 13 Topic - Methods, History, Theory

Abstract

Developmental research has revealed the critical importance of early skills and experiences, including during the neonatal period, for later development. Tools that can translate laboratory-based behavioral paradigms to measure behavior longitudinally in large samples of newborns remain lacking. This limits options for large-scale behavioral studies in this age range, which are needed to develop large-scale assessments and interventions that build upon laboratory findings. We address this gap by examining the feasibility of eliciting and measuring a neonatal behavior, neonatal mimicry, commonly studied in tightly controlled laboratory settings, in a caregiver-led online paradigm. Our between-subjects design focuses on two of the most commonly studied neonatal gestures in the context of mimicry, tongue protrusion (TP) and mouth opening (MO). Caregivers and their newborns were videotaped, via the online experiment platform Lookit, as caregivers modelled either gesture to their newborns (Figure 1). To assess data quality in this online paradigm, coders separately coded the pilot data (N = 12, all sessions) obtained from Lookit for total usable time, that is, the total task time with valid video usable to code TP and MO. In this pilot webcam data, our coders demonstrated a high degree of inter-rater reliability for coding total usable time, r(5) = 0.95, p < 0.001 (Figure 2, left). Usable time is defined by the infant’s face being in view, eyes open, and not fussing or bouncing for an extended period of time (i.e., > 3 sec.). Based on the coded videos (N = 7, codable videos), 71% (N=5) met laboratory-based inclusion criteria (total usable time > 90 sec.); the remaining 5 sessions did not contribute a task video (ended prior to starting the test trial, N = 4, technical issues with webcam, N = 1) (Figure 2, right). Participants took a median time of 12.20-min. to complete the task for included sessions (12.86-min., all sessions), in line with our projected task duration of 15-min. This initial analysis demonstrates the feasibility of measuring neonatal mimicry of caregivers in the home using this paradigm. Coding of newborns’ gesture frequencies in the pilot data is underway to determine whether TP and MO neonatal gestures may be coded reliably from webcam videos using laboratory-based standards, prior to pre-registration and launching of experimental data collection. Pre-registered analyses of the to-be-collected experimental data will examine the experimental phase, using a Bayesian hierarchical log-linear model, to test whether the frequency of MOs for neonates in the MO condition increased when caregivers modelled MO and vice versa for the TP condition. Models will be estimated separately for MO and TP. We expect that this will validate our novel online paradigm for measuring and eliciting neonatal mimicry of caregivers in the home environment, providing a conclusive estimate for the prevalence of neonatal mimicry at the group level. We estimate that a sample size of approximately N = 290 (145 per condition) will be necessary to ultimately reach a high standard of evidence; in the context of the current submission we plan to report on the interim data set to be obtained.

Group Authors

N/A

Authors

• Katherine E Casey, American University
  Presenting Author

• Kimberly Scott, Massachusetts Institute of Technology (MIT)
  Non-Presenting Author

• Kira Ashton, American University
  Non-Presenting Author

• Jeff Gill, American University
  Non-Presenting Author

• Elizabeth A. Simpson, University of Miami
  Non-Presenting Author

• Laurie Bayet, American University
  Non-Presenting Author