Variation in Everyday Caregiver Sensitivity Predicts Infants’ Real-Time Physiological Regulation

• In Event: Leveraging Wearable Sensors to Access Processes of Children’s Everyday Behavioral and Physiological Regulation

Abstract

Caregiver sensitivity to distress is a well-established predictor of infant social-emotional development and contingent, appropriate responses to infant distress are thought to support the regulation of infants’ distress in real-time (Leerkes & Zhou, 2018). However, almost all work examining caregiver sensitivity to distress occurs in structured laboratory paradigms, which restrict variability associated with “everyday caregiving” where caregivers may juggle other needs and responsibilities at home, including care of other children. In past work, we have found substantial within-person variation in mothers’ sensitivity to infant distress over the course of the day (Authors, 2022).

In the current paper, we examine the effects of variation in everyday caregiver sensitivity to distress on infant real-time behavioral and physiological regulation. Drawing from Porges Polyvagal theory (Porges, 2007), we hypothesize that within-person, on occasions that caregivers are more sensitive than average, their infants will show greater decreases in respiratory sinus arrhythmia, an established marker of adaptive regulation of distress (Sacrey, et al., 2021).

N = 87 families shared 72+ hours of infant-worn synchronized audio and heart rate sensors collected over the course of a week. N=38 24-h recordings, each from a distinct infant, were selected for detailed annotation. Research assistants annotated all infant fussing and crying in LENA audio recordings, grouping individual instances of distress occurring within five minutes into a single episode (ICC=0.85). A second team annotated caregiver sensitivity to distress for each identified distress episode (ICC=0.83). To examine infant physiological regulation we examined dynamic changes in Respiratory Sinus Arrhythmia surrounding distress episodes identified in audio recordings. RSA timeseries were derived for each episode of distress using a 15s moving window technique to create a 5Hz timeseries of RSA (Abney, et al., 2021). RSA timeseries were time-locked to the onset and offset of individual distress episodes. RSA change was calculated by subtracting the RSA value in the sample prior to distress onset or offset from the RSA timeseries.

On average, infants had 12.3 (range 1-26) episodes of distress in 24h, ranging from 1-58 minutes in duration (M=5.5min, SD=2.3min). 17% of responses to distress were from individuals other than the mom. Excluding all non-mom responses, multilevel models indicated that rankings of everyday maternal sensitivity to distress were highly variable across individual instances of distress, with 63.5% of differences attributable to within-person variance (see Figure 1). Preliminary analyses compared RSA around distress for a subset of infants (n=17, expected N=38 for March 2023) using a median split of average caregiver sensitivity. We found that lower average caregiver sensitivity was associated with less efficient physiological regulatory strategies (decreases for 18.4/33s vs 28.2/33.15s under confidence interval, p<0.05). Building off of these analyses, we plan to use multilevel models to test how deviations from within-caregiver mean sensitivity predict variations in the relative change in RSA following the onset and offset of infant distress.

Authors

• Anna Madden-Rusnak, University of Texas at Austin
  Presenting Author

• Megan Micheletti, University of Texas at Austin

• Brandon Thint

• Priyanka Khante, University of Texas at Austin

• Xuewen Yao

• Kaya de Barbaro, University of Texas at Austin