The effect of stimulus and model characteristics on childhood vicarious fear learning and unlearning

Rachman (1977) suggested that fears may be acquired via three distinct pathways: direct traumatic experience, verbal information received from others, and vicarious learning. In vicarious fear learning an individual learns from another individual by observing their response to a stimulus or situation. Mineka and Cook (1993) showed that in monkeys this observed fear response behaves as an unconditioned stimulus (US). Thus vicarious learning is likely to be similar, in a procedural sense, to classical conditioning and the success of learning is determined by the relative strength of these conditioned stimulus-unconditioned stimulus (CS-US) associations. Recent studies suggest that vicarious learning plays a role in the development of childhood fears (Askew & Field, 2007; Gerull & Rapee, 2002). However, little is known about the mechanisms underlying this type of learning and the factors that affect it. Associative learning research shows that the salience of the US (a measure of its effectiveness) is critical for learning, if the US is more salient, learning should be more effective. Therefore, the first group of experiments in this thesis manipulated factors expected to affect US salience and hence vicarious fear learning. Experiment 1 considered the importance of the relatedness of observers and models for US salience, specifically whether greater increases in children's fear-related responses are observed when fear responses are modelled by children's mothers compared to when they are modelled by strangers. It also looked at whether fear beliefs acquired vicariously can be unlearned via vicarious counterconditioning and whether the type of model influenced this. Results showed that children's fear responses for animals increased following fear-related vicarious learning and decreased following positive learning. Similarly, counterconditioning led to unlearning of these responses. Model type did not affect vicarious learning or subsequent vicarious counterconditioning. Experiment 2 looked at another factor expected to affect US salience: the age of the model relative to the observer. This experiment examined the relative effects of same-age peer modelling and adult modelling in changing children's fear-related responses. Results were nearly identical to the first experiment: learning and unlearning were similar irrespective of whether the model was a peer or an adult. Experiment 3 examined what effect the richness of the information source has on US salience using moving models (on film) compared to still (photographic) images. No significant effect of the threat-relevant vicarious film on children's fear beliefs and attentional bias was found. The second group of experiments concentrate on characteristics of the CS. Research with adults suggests that when fear-relevant stimuli are used, fear learning will generally be greater (produce a larger conditioned response), occur more rapidly (in fewer trials) and will be more robust (demonstrate superior resistance to extinction) than for fear-irrelevant stimuli (see e.g. Öhman & Mineka, 2001). However, Askew, Dunne, Özdil, Reynolds, and Field (2013) showed that vicariously acquired conditioned responses for fear-relevant stimuli were not significantly greater than those for fear-irrelevant stimuli in children (6 to 11 years). A further two experiments here examined the speed and robustness of vicariously learned responses for fear-relevant stimuli. Experiment 4 manipulated the number of CS-US pairings children saw and compared vicarious learning for two stimuli of differing fear-relevance. Results found increased fear-related responses for both types of stimuli: the effect was not affected by the fear-relevance of the stimuli, nor the number of CS- US pairings that children were exposed to. Fear beliefs were still found to be raised at follow-up one week later, and no . evidence was found for more robust learning for more fear-relevant stimuli. The final experiment, Experiment 5, looked at robustness in more detail. The resistance of learnt responses to extinction was compared in two stimuli of varying fear- -. relevance (animals and flowers). Significant increases in fear-related beliefs and avoidance for animals and flowers were found again, but there was no significant extinction effect observed for avoidance preferences for either CS type. Unexpectedly, significantly greater extinction was observed with fear beliefs for animals compared to flowers. In summary, the first set of experiments showed that in VL with children, the salience of the US is either not easy to manipulate or, more likely, US salience has little effect on fear learning. The second set of experiments demonstrated that the fear-relevance of the stimulus is also not important for children in VL. These findings may have an evolutionary explanation: for children of this age, acquiring knowledge of a potential threat and acting upon it may be more important in terms of survival than the salience/relevance of the US or CS that facilitates such acquisition