(PS4-A20) Creating a 'Fingerprint' to Predict Early Childhood Psychopathology Using Naturalistic Cry Audio

Background:

While tantrums and irritable mood are normative misbehaviors of infancy and early childhood, when fraught and/or dysregulated they may also be indicators of clinical concern (Wakschlag et al., 2018, 2019). Thus, it is important to distinguish early risk markers for normal vs. abnormal irritability (Cook et al., 2019; Krogh-Jesperson et al., 2022; Van den Akker et al., 2022; Hoyniak et al., 2022). The current study examines in vivo, naturally occurring infant vocalizations as a predictor of later psychopathology.

To date extant research on infant irritability primarily uses retrospective parent-report. New wearable infant recording technology, like the Language Environment Analysis System (LENA; Gilkerson et al., 2017), allows researchers to examine associations between real-time irritable behaviors and later psychopathology, potentially creating profiles, or ‘fingerprints’ of clinical risk as early as infancy.

Here, we examine the associations between frequency, duration, and intensity of 12-month-old infant vocalizations, and 12-month irritability and 36-month psychopathology. We hypothesize that greater 1. Frequency and 2. Duration of irritable vocalizations will be positively associated with irritability at 12 months and psychopathology symptoms at 36 months. We also predict that lower intensity vocalizations (whine/fuss) will be negatively associated, and higher intensity vocalizations (cry/scream) positively associated, with irritability at 12 months and psychopathology symptoms at 36 months.

Methods:

Our sample includes 100 racially and economically diverse families from a longitudinal study oversampled for irritability. Infants wore LENAs to capture day-long audio recordings at 12 months (Gilkerson et al., 2017). Irritability at 12 months was assessed via the Multidimensional Profile of Disruptive Behavior – Infant-Toddler Version (MAP-DB-IT; Wakschlag et al., 2012), and psychopathology symptoms at 36 months via the Preschool Age Psychiatric Assessment (PAPA; Egger at al., 2004). Using ELAN annotation software (Zheng et al., 2022), trained research assistants coded randomly selected 10-min increments of the recordings (1,000 minutes total) for irritable vocalizations. We double-coded n = 20 recordings for reliability. We will calculate frequency, duration, and proportion of intensity of vocalizations for each recording.

Results:

Descriptive analyses revealed a 12-month irritability z score of M = .53 (SD = .62). Pearson correlation coefficients will be used to examine associations between frequency, duration, and intensity of naturalistic vocalizations and irritability at 12 months and psychopathology symptoms at 36 months. Analyses will be complete by the 2023 ABCT Annual Convention.

Conclusion:

Naturalistic assessment of irritable vocalizations may allow researchers to create more accurate ‘fingerprints’ of clinical risk. Early, accurate identification of concerning irritability would facilitate innovative prevention initiatives, or earlier interventions, as a cost-effective, targeted strategy for children who may be at risk (Wakschlag et al., 2019).