Defining multi-dimensional associations in fear conditioning: implications for abnormal fear responses

Abstract

Post-Traumatic Stress Disorder (PTSD) and anxiety disorders are related to the development of an abnormal fear response that is triggered in situations where a fear response is not appropriate (McSweeney et al., 2014). The current Pavlovian model of fear conditioning contends that animals learn a simple prediction between a cue or context and an aversive stimulus (e.g., footshock). However, in conditions such as PTSD and anxiety the fear response is activated in situations that are not consistent with the prediction, indicating that something other than just a simple prediction may be learned in context fear conditioning. This would seem consistent with research indicating that motivational or affective components are also learned in Pavlovian appetitive conditioning (Berridge, 2018). In this study we aim to lay the groundwork for identifying motivational or affective associative processes in Pavlovian fear conditioning. In addition, our research examined differences between males and females in the acquisition of learned associations, and how these differences may lead to varied generalizations between contexts. We used Pavlovian fear context conditioning to elicit predictions using both high-prediction-training (continuous reinforcement) and low-prediction training (partial reinforcement). A blocking protocol was used to test for memory retrieval and to evaluate the strength of the prediction. Tone testing was conducted to ascertain if there are differences in freezing levels between males and females. In our results, we observed minor differences in blocking levels, and differences in context generalization between males and females, supporting our hypothesis that there are differences in learning between the sexes. Our main finding - that while both high-prediction training and low-prediction training lead to strong predictions only high-prediction training blocks new learning - is evidence that prediction is not sufficient to explain all fear conditioning. This supports our contention that reframing fear conditioning toward a model based on a multi-dimensional set of associations could have a major impact on the diagnosis and treatment of PTSD and anxiety disorders.