Fear generalization, in which conditioned fear responses generalize or spread to related stimuli, is a defining feature of anxiety disorders. The behavioral consequences of maladaptive fear generalization are that aversive experiences with one stimulus or event may lead one to regard other cues or situations as potential threats that should be avoided, despite variations in physical form. Theoretical and empirical interest in the generalization of conditioned learning dates to the earliest research on classical conditioning in nonhumans. Recently, there has been renewed focus on fear generalization in humans due in part to its explanatory power in characterizing disorders of fear and anxiety. Here, we review existing behavioral and neuroimaging empirical research on the perceptual and non-perceptual (conceptual and symbolic) generalization of fear and avoidance in healthy humans and patients with anxiety disorders. The clinical implications of this research for understanding the etiology and treatment of anxiety is considered and directions for future research described.

Adult male subjects saw a sexual film clip paired with a nonsense syllable (C1). Similarly, an emotionally neutral film clip was paired with a second nonsense syllable (C3). Responses to the nonsense syllables were recorded as skin resistance responses. Subjects were also trained in a series of related conditional discriminations, using the C1 and C3 stimuli, from which the following equivalence relations were predicted; A1-B1-C1, A2-B2-C2, and A3-B3-C3. Some subjects were given matching-to-sample (equivalence) tests after the conditional discrimination training (Experiment 1), whereas others were not (Experiment 2). Subjects were tested for a transformation of eliciting functions by presenting the A1 and A3 stimuli, which were related through equivalence to C1 and C3, respectively. Five of the 6 subjects who showed significantly greater conditioned responses to C1 than to C3, also showed significantly greater skin resistance responses to A1 than to A3. Two additional subjects demonstrated a transformation of an eliciting stimulus function in accordance with five-member equivalence relations (Experiment 3), and another 5 subjects demonstrated similar effects in accordance with the relations of sameness and opposition (Experiment 4).

Conflict of interest: Ian Tyndall declares that he has no conflict of interest. Daniel Waldeck declares that he has no conflict of interest. Robert Whelan declares that he has no conflict of interest. Bryan Roche declares that he has no conflict of interest. David Dawson declares that he has no conflict of interest. Luca Pancani declares that he has no conflict of interest.

Despite the central status of avoidance in explaining the etiology and maintenance of anxiety disorders, surprisingly little behavioral research has been conducted on human avoidance. In the present paper, first we provide a brief review of the empirical literature on avoidance. Next, we describe the implications of research on derived relational responding and the transformation of functions for a contemporary behavioral account of avoidance, before providing several illustrative research examples of laboratory-based analogues of key clinical treatment processes. Finally, we suggest some challenges and opportunities that lie ahead for behavioral research on anxiety and avoidance.

No abstract