Poster #122 - Ambulatory Assessment of Infant Negative Affectivity: Comparison to Gold-Standard Temperament Assessments

• In Event: 1-169 - Poster Session 04
  In Poster Session: PS 04 Section - Methods, History, Theory

Integrative Statement

The current gold-standard methods for measurement of infant temperament include parent questionnaires and short observational assessments, both of which have their own strengths and weaknesses (Stifter & Dollar, 2016). With the current ubiquity of mobile and wearable sensors, ambulatory capture of daily activity holds promise as an objective measure of temperament that could more robustly characterize behaviors and reactions over extended periods of time. Additionally, the possibility to automatically detect markers of ambulatory activity that speak to temperament can facilitate more dense longitudinal sampling and ultimately better data to address questions of why and how individual differences in temperament develop.

In the current poster, we present data assessing relationship between detected episodes of infants’ distress, from 24-72 hours of natural daily activity and gold standard measures of observed behavioral and physiological reactivity (Garcia Coll et al. , 1988) as well as maternal-reported negative affectivity (Garthstein Rothbart et al. 2003).

Infants (N= 8, expected N= 35; age: 6 weeks to 9 months) participated in a structured home session followed by 24h of naturalistic recording. The structured home session included a sensory challenge task that assessed infants’ reactions to standardized presentation of five different stimuli (hair brushed, tissue on face, playground noises, lemon juice and noisy toys), following Garcia Coll et al and Kagan. Behavioral reactivity was rated on a scale of 0 (None) to 4 (Extreme), considering infants’ vocal, motor (subtle and gross) and facial reactions, also following Garcia Coll et al. (1988; 80% interrater accuracy). Physiological reactivity was assessed by change in heartrate from a five-minute baseline period to the sensory challenge. Preliminary analyses indicate that Mean HR during the challenge task (M= 138.47 bpm, SD= 7.67) was significantly higher than the baseline (M= 117.39bpm; SD= 6.39; p<.001 via t-test), suggesting that our challenge protocol is a valid manipulation of infant arousal state.

During the 24-72h naturalistic home recordings, infants wore motion and physiological sensors and all sounds occurring within 6-10 feet of the infant were recorded by LENA© recorders worn by infants in a fitted vest. For multi-day sessions, multiple recorders were used. We developed and tested an algorithm to create distress episodes from standard LENA outputs of individual infant negative vocalizations. Preliminary analyses of LENA data (N=8 16-hour recordings) indicate that the validity of infant distress episodes is increased by 20% (88.44% vs. 66.7%) by removing crying episodes with exclusively short (>2sec) negative vocalizations.

Preliminary analyses with eight participants indicated a heterogenous pattern of correlations between traditional/ gold-standard assessments of observed (behavioral and physiological) and reported measures of reactivity (range: r= .02 -.56; see Figure 1). Correlations between measures of automatically estimated overall daily crying appear to be stronger with observed (behavioral and physiological) measures, indicating that ambulatory measures may be able to stand in for these assessments while complementing maternal-report measures. Space permitting we will also explore other dimensions of temperament assessed by the IBQ and in the lab with our 24h data, such as the duration of individual cries as a measure of soothability and or the total daily motion as a measure of activity level.

Authors

• Kaya de Barbaro, The University of Texas at Austin
  Presenting Author

• Lara Andres, Texas A&M University
  Non-Presenting Author

• Grace Chang, The University of Texas at Austin
  Non-Presenting Author

• Mckensey Johnson, The University of Texas at Austin
  Non-Presenting Author