Capturing Temporal Dynamics of Fear Behaviors on a Second by Second Basis

• In Event: 3-109 - Understanding Processes underlying Children's Emotional Development and Emotion Regulation Using Intensive Longitudinal Approaches

Integrative Statement

Children with Dysregulated Fear (DF) express high levels of fear that do not match the situation. It is important to identify children with DF early and accurately because they are at increased risk for anxiety (Buss et al., 2013). Previous work has focused on the total amount of fear by using composites across time. However, the temporal dynamics of fear expression might offer novel insights in the identification of children at risk and inform prevention programs to target specific behaviors that in the moment will perpetuate anxious behavior.

The sample consists of 124 children (61 girls, Mage = 24.43 months, SDage = .47). The children participated in a high-fear task where they were in a room with the caregiver and a giant toy spider and a low-fear task in a room with the caregiver and a puppet show, designed to be novel but not fear evoking. Each behavioral task was recorded and later coded second by second by a team of trained and reliable coders. The three behaviors coded were facial fear, bodily fear, and proximity to caregiver.

Data were modeled using a Hidden Markov Model, which approximates children’s latent underlying fear states and models their state transitions over time. Results revealed that children's behavior is best represented as transitions among six behavioral states, characterized by variation in fearfulness and proximity to caregiver (CG) (see Table 1). Transitions between states differed significantly between the two tasks (Chi-square likelihood ratio test: (30) =12643.99, p<.001.) Children in the spider task showed overall more probability of changing state, with more transitions between high-fear and low-fear states, suggesting higher volatility in fear in this condition. Children in the spider condition also showed higher probabilities to transition directly back and forth between no-fear/at CG and high-fear/away, suggesting that the caregiver may be acting more strongly as a regulating force. These underlying states were analyzed using sequence clustering to identify groups with similar dynamic trajectories through the states. Results showed that four clusters fit the data best in both high and low fear tasks as well as across tasks. Across-task clusters included “external regulators” (children 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 “DF” (fear/at caregiver state regardless of task). In the high fear task, there were three fearful clusters, “high”, “moderate”, and “low” fear, as well as the “low reactive” cluster. In the low fear task, the four clusters that emerged were “fearful explorers”, “Low reactive”, “DF”, and “internal regulators” (where children spend much of their time far from caregiver and showing no fear, but still showing more transitions between fear, low fear, and no fear states). High fear and low reactive clusters were consistently apparent across high and low fear tasks, demonstrating DF as a consistent fearful group. Results also suggested both internal (far from caregiver but still showing low levels of fear) and external (fear low only near caregiver) regulation strategies.

Authors

• Elizabeth A Shewark, The Pennsylvania State University
  Presenting Author

• Timothy R Brick, The Pennsylvania State University
  Non-Presenting Author

• Kristin A Buss, The Pennsylvania State University
  Non-Presenting Author