Stimulus variability during extinction training does not improve the generalization of fear extinction learning

Fear extinction procedures serve as a laboratory model for a learning process involved in exposure treatment for anxiety disorders. Clinically, exposure is typically conducted with generalization stimuli (GSs), because originally acquired fear stimuli are inaccessible. Experimental studies, however, show limited generalization of extinction when GSs are used in extinction training (i.e., GS extinction), which may represent a risk for return of fear. Stimulus variability may overcome limited extinction generalization. This study examined whether using multiple, perceptually similar GSs during fear extinction enhances GS extinction generalization. A healthy sample (N = 120) underwent a two-day fear conditioning procedure. During extinction, participants repeatedly experienced either a single GS, multiple GSs, or the original conditioned stimulus (CS+). Later, threat expectancy and skin conductance responses to novel stimuli and the CS+ were measured during three test phases. Across groups, return of fear on the second day was evident for the CS+. Contrary to our hypotheses, no group differences during test phases were apparent, suggesting comparable extinction generalization independent of extinction training stimuli. This finding was replicated in an additional web-based study. Results point to questions about the influence of (missing) awareness of perceptual differences and its interplay with higher-order cognitive processes for extinction generalization.