Neurobehavioral Mechanisms Supporting the Generalization of Learned Fear in Humans

Abstract

An inescapable component to survival in a dynamic environment is detecting and reacting to signals of danger. One of the most elegant processes animals possess to handle this complex task is classical conditioning, wherein stimuli associated with an aversive event acquire the capacity to elicit defensive behaviors. This process helps ensure quick reactions prior to the occurrence of an imminent threat. A problem of living in a dynamic environment, however, is that reliable signals of danger are rarely re-encountered in the exact same form from one situation to the next. Thus, to be truly adaptive it is imperative for defensive responses to extend beyond a specific instance towards other exemplars that might portend the same negative outcome. While the phenomenon of stimulus generalization was recognized in the earliest studies of conditioning from Pavlov's laboratory, a century of conditioning research has not resolved how humans and other animals actually meet this challenge. The research presented herein employs a combination of psychophysiological and functional imaging methods to examine how humans recruit neurocognitive systems to determine what stimuli do (and do not) pose a threat. Results show that human fear generalization is a complex phenomenon affected by the perceptual and conceptual nature of the stimulus. Brain regions and functional networks involved in fear generalization comprise cortical areas involved in coding the representation of conditioned stimuli and subcortical regions involved conditioned learning and the production of behavioral responses, most notably the amygdala. These results reveal the importance of stimulus-specific factors in fear learning and generalization, provide support for anatomically constrained models of fear generalization, and contribute to the development of model systems of fear generalization processes in human anxiety disorders.