A flat month and a prediction that refuses to move
- N = 77 adults with social anxiety disorder or panic disorder completed a nine-session exposure protocol. 41 were randomized to Inhibitory Retrieval exposure, built on extinction learning, and 36 to Habituation exposure, aimed at fear reduction.
- In the Inhibitory Retrieval condition, participants whose positive affect was low relative to their own mean reduced their expectancy of the aversive outcome less when that expectancy was high before the exposure (p = .029).
- In the same condition, participants whose negative affect was high relative to other participants also reduced high pre-exposure expectancies less (p < .001).
- In the Habituation condition, neither positive nor negative affect moderated the change from pre- to post-exposure fear. Affect was measured with the Positive and Negative Affect Schedule (past month) at baseline, mid-treatment and post-treatment.
In exposure built on expectancy violation, the client names what they think will happen, does the task, and rates the prediction again. The gap between the two ratings is where the learning is meant to sit. A secondary analysis of a randomized trial by a team that includes Michelle Craske of UCLA, published in the Journal of Anxiety Disorders, asks whether the client's mood changes the size of that gap.
Two protocols and a mood questionnaire
The 77 participants had social anxiety disorder or panic disorder and went through nine sessions. In the Inhibitory Retrieval arm, 41 people rated the expectancy of an aversive outcome before and after every exposure. In the Habituation arm, 36 people rated their fear before and after instead. Everyone filled in the Positive and Negative Affect Schedule for the past month three times: at baseline, mid-treatment and post-treatment. The authors compared each person with their own average and with the other participants.
In the Inhibitory Retrieval arm, both kinds of affect mattered when the starting expectancy was high. Positive affect below the person's own mean went with a smaller drop in expectancy (p = .029). So did negative affect above that of other participants (p < .001). In the Habituation arm, affect did not moderate fear change. The authors suggest that exposure models could add strategies that raise positive affect.
Asking about the good weeks too
Mood here shapes how much the client takes away from a prediction that did not come true. The usual pre-exposure check looks for distress. These data add the other side: a client who is not especially anxious, but flatter than they were a month ago, may finish a well-run exposure with a high prediction barely shifted.
Before an exposure based on expectancy violation, ask about the recent weeks, including what has been enjoyable and how that compares with the client's usual. Low positive affect is a signal in its own right, not only high distress. When it is there, spend more time on the post-exposure rating and on what the client concludes from it.
Low positive affect is a signal in its own right before an expectancy-violation exposure, not only high distress.
This is a secondary analysis of a randomized trial with 77 participants and two diagnoses. Affect was rated for the past month at three time points, not immediately before each exposure, and the moderation results describe associations.