Capturing temporal dynamics of fear behaviors on a moment-to-moment basis.

Identifying patterns of fearful behaviors early and accurately is essential to identify children who may be at increased risk for psychopathology. Previous work focused on the total amount of fear by using composites across time. However, considering the temporal dynamics of fear expression might offer novel insights into the identification of children at risk. One hundred and twenty-five toddlers participated in high- and low-fear tasks. Data were modeled using a novel two-step approach. First, a hidden Markov model estimated latent fear states and transitions across states over time. Results revealed children's behavior was best represented by six behavioral states. Next, these states were analyzed using sequence clustering to identify groups of children with similar dynamic trajectories through the states. A four-cluster solution found groups of children varied in fear response and regulation process: "external regulators" (using the caregiver as a regulation tool), "low reactive" (low reaction to stimulus), "fearful explorers" (managing their own internal state with minimal assistance from the caregiver), and "high fear" (fearful/at-caregiver state regardless of task). The combination of analytic tools enabled fine-grained examination of the processes of fearful temperament. These insights may help prevention programs target behaviors that perpetuate anxious behavior in the moment.