The validity of human avoidance paradigms

Excessive avoidance towards innocuous cues is a central diagnostic criterion across anxiety-related disorders. Relevant laboratory paradigms typically include instrumental learning procedures, where the presentation of an aversive cue (e.g., a shock) can be prevented by executing an experimenter-defined response (e.g., a button press) during the presentation of a warning cue (e.g., a square). Despite the popularity of these paradigms, there is no evaluation of how well the experimental findings of conditioned avoidance extend to maladaptive avoidance, or whether findings from animal studies could be informative for human studies. Here, we present a validation of the conditioned avoidance paradigm. We show that although this procedure meets the majority of the tested validity criteria (i.e. face, construct, predictive, and diagnostic validity), it also faces a number of challenges, including the non-consideration of individual differences in learning or the use of procedures that cannot be easily translated to clinical settings. For meeting these challenges, we suggest extensions of the paradigm including the test of individual differences by using ambiguous stimuli as well as the use of virtual reality procedures. Our main conclusion is that despite the significant knowledge provided in conditioned avoidance paradigms, their expansion will allow reaching more theoretical and clinical insights.