Fear one, fear them all: A systematic review and meta-analysis of fear generalization in pathological anxiety.

It is a defining feature of anxiety disorders that fear is elicited by a circumscribed class of stimuli rather than by only one specific exemplar of that class. Therefore, fear generalization, a mechanism by which associative fear extends from one conditioned stimulus to similar cues, has been central to theories on anxiety. Yet, experimental evidence for the link between generalization and pathological anxiety, as well as its moderators, has not been formally integrated. This systematic review and meta-analysis of empirical findings clarifies the relationship between fear generalization and pathological anxiety. In conclusion, enhanced fear generalization is associated with several anxiety disorders and stress-related disorders, which is supported statistically by a small, but robust effect size of g = 0.44 for risk ratings as an index of fear generalization. However, empirical results are inconsistent across disorders and they rarely allow for conclusions on their causality in the disorders' etiology. Therefore, based on theoretical considerations, we recommend directions for intensified research, especially on the causal relationship between overgeneralization and pathological fear.

Hear it, fear it: Fear generalizes from conditioned pictures to semantically related sounds.

Fear generalization is thought to be an important mechanism in the acquisition and maintenance of anxiety disorders. Previous studies have investigated fear generalization within one sensory modality - mainly within the visual domain. However, a growing body of evidence shows that emotional information is processed in more than one sensory modality. Based on network theories, we expected that fear may also generalize from stimuli in one sensory modality to another. To test our hypothesis, 42 participants underwent a differential conditioning paradigm, during which pictures were either presented with (vCS+) or without (vCS-) an aversive electric stimulus. After the acquisition phase, generalization was tested in the crossmodal group (n = 21) by presenting sounds which were semantically congruent to the visual vCS+ (i.e., the aGS+) or the vCS- (i.e., the aGS-). As a control, the unimodal group (n = 21) saw the pictures again. For the crossmodal group, we could show that US expectancy ratings generalized from conditioned pictures (vCS+) to semantically related sounds (aGS+). Moreover, when the vCS+ was presented during extinction, fear of the aGS+ extinguished, whereas extinction training with the aGS+ was found to be less effective for the vCS+. The findings are relevant for crossmodal fear acquisition and exposure therapy.