March 25, 2013

Experimental Modification of Interpretation Bias about Animal Fear in Young Children: Effects on Cognition, Avoidance Behavior, Anxiety Vulnerability, and Physiological Responding

Lester, Field, & Muris, 2011

Research Question

  •       How does experimental modification of children’s interpretation bias affect cognition, behavioral avoidance, anxiety vulnerability, and physiological responding?
Introduction and Literature Review

  •         Child anxiety is associated with the tendency to interpret ambiguous information in a threatening manner
  •       Adult studies using feedback-learning paradigms (CBM-I) have demonstrated that it is possible to induce positive and negative interpretation biases, and that negative biases are causally linked with anxiety vulnerability
o   Participants are reinforced for consistently disambiguating scenarios in a positive or negative way

  •        CBM-I procedures have been developed that are more suitable for use with children
o   Important because childhood represents a “hot” developmental period when negative cognitions (presumably) and anxiety develop
o   Successful in changing cognition, behavior, and anxiety vulnerability, but not state anxiety

  •        Children who had been given negative information about an ambiguous animal showed higher heart rates when asked to approach a box they thought contained the animal, compared to a control animal

  •         Children in the negative modification condition will interpret future ambiguous information in a negative manner, and vice versa for the positive condition
  •         Negative modification will lead to an increase in anxiety across the training, positive modification will lead to a decrease in anxiety
  •         Compared to those in the negative training, kids in the positive training condition will demonstrate less avoidance behavior, anxiety vulnerability, and physiological responding
·         Participants
o   67 children, 6-11 years of age (29 girls, 38 boys)
o   Randomized to positive animal modification (34) or negative animal modification (33)
·         Procedure
o   Mean heart rate measured
o   Completed STAI-C and FSSC-R
o   Ambiguous vignettes
§  18 vignettes counterbalanced so that 9 are administered pre-modification and the other 9 post-modification (to test the effects of the modification)
§  5 sentences long and children indicate at the end of each sentence how scary they thought the ending of the story would be (1 to 7 Likert scale)
§  Ex. “It is night time and you are in your bedroom. You hear a noise coming from outside. The noise gets louder. You open your curtains to take a look outside. You see a cuscus outside your window.”
o   Visual analogue mood scales
§  Children were asked to rate from 0 to 100 how nervous they felt immediately pre- and post-modification
o   Bias modification
§  Programmed in E-Prime
§  30 ambiguous scenarios involving 3 novel Australian marsupials
§  Children are asked to imagine themselves in each situation and decide how each story might continue (by choosing either a positive or negative option)
§  After choosing, kids were provided with feedback contingent on their condition (either “Good” or “Wrong”)
§  Ex. “You are in a field when you spot a quoll looking around for something to eat. There are lots of little baby mice scurrying around the field. You also notice some bushes full of juicy berries. You wonder what the quoll will like to eat.”
·         “Oh no!! They quoll is gobbling up the baby mice. You can hear his sharp teeth crunching their bones” or “ The hungry quoll runs over to the buses and starts munching on the tasty berries and leaves”
o   Visual analogue mood scales
o   Ambiguous vignettes
o   Behavioral Approach Task (BAT)
§  Anxiety rating
§  Covered pet carrier with a slit on one end, soft toy, hay, and rustling sounds – asked to approach and pet the animal
§  Anxiety rating
o   Debriefing and optional positive bias modification

  •         Groups were comparable on trait anxiety, fear scores, and gender
  •         The training worked: kids in the positive condition learned to pick the positive outcomes and kids in the negative condition learned to pick the negative outcomes
  •         Threat biases decreased significantly after the positive modification, but increased significantly after the negative modification
o   But the negative group was more anxious pre-modification

  •         Increase in anxiety across the negative modification, no change in anxiety across the positive modification
  •       BAT results may have been affected by the fact that kids in the negative training condition were more anxious pre-BAT
o   Children in the negative group took longer to approach the pet carrier than children in the positive group
o   Odds of not putting their hand in the box was 3.52 times higher for those in the negative condition
o   Training did not have an effect on anxiety vulnerability or physiological responding to the task

  •         Interpretation biases can be successfully modified in young children – the training worked and had associated effects on interpretation bias, state anxiety, and behavioral avoidance, but no effects on anxiety vulnerability or physiological responding
  •         Effects of positive modification stronger than negative modification
  •         No way to entirely tease out if induced biases caused by changes in state anxiety or the training itself, but past research suggests the latter
o   Inconsistent literature on modification affecting state anxiety

  •         Mood differences pre-BAT may help to explain group differences in behavioral avoidance
  •          Active generation of interpretations, longer training, or assessment of measures post-BAT may be required to see group effects on anxiety vulnerability and physiological responding
o   Or a more emotionally salient task may be necessary to evoke responses in all three components of anxiety

  •          Limitations: nonclinical sample, wide age range, different methods from those used with adults
  •          Future clinical applications highlighted


  •          Mean heart rate was calculated over a 15-second period during the BAT and the task was stopped if kids didn’t put their hand inside in that amount of time. Is reaction time the best way to measure behavioral avoidance? Any problems with this approach to measuring heart rate? Any ideas on how to measure behavioral avoidance or physiological responding differently?
  •          Anxiety significantly increased across the negative training, but non-significantly decreased across the positive training. Any thoughts on why this might be?
  •          How might you design the study differently so that you could tease apart the effects of cognition and emotion on behavioral avoidance during the BAT?
  •         The authors argue that they might have been more likely to find group differences on anxiety vulnerability and physiological responding if they had assessed these measures post-BAT. Do you agree? What about a prolonged pre-BAT period to assess anticipatory anxiety?
  •          Interpretation training was effective in changing behavioral avoidance, but not physiological responding or subjective ratings of anxiety. How can we make sense of this given what we know about CBT for child anxiety?
  •         What do you think about using trainings such as these as an adjunct to CBT for child anxiety? Are there childhood anxiety disorders you think might be particularly amenable?
  •          What differences, if any, might you expect in a sample of clinically anxious children?


  1. Just a quick response to this article -- I don't know much about CBM, but how much do we think the cognitions of these kids really changed? I can see them "learning" to give the correct response, but not actually believing it. I guess the behavioral task gets around that somewhat, but I'm wondering if it has a lot to do with the kids trying to please the experimenters (something along the lines of, "they just told me things things are safe and really want me to touch the animal, so I should do it"). Any thoughts?

  2. Excellent point as this is one of the most common critiques out there against CBM. I agree that there could certainly be some demand characteristics at play. This could potentially explain the lack of group differences in physiological responding, as one might expect kids in the positive condition to still be physiologically aroused if they didn't actually believe the animal was safe and were only approaching it because they knew they were "supposed to." On the other hand, this wouldn't really fit with the finding that there was no change in state anxiety across the positive modification.

    Would you be more convinced if the effect was shown to be more generalized, for example, showing that the positively trained group was more open to approaching other novel animals not discussed in the training? Are there other outcomes measures that you would find more convincing?

  3. Thanks for your reply, Kristy. Yeah, I think I'd be more convinced if the behavioral approach task generalized to other types of animals, since the kids' thinking at that age may be concrete enough that they might not think what they've been told applies to all animals. Also, perhaps if the CBM task were more active - like the example they give for adult versions where adults have to complete a word or sentence - that would be more convincing, since that seems more akin to actual cognitive bias.