(PS9-B43) Core Fears and Infinite Repetitions: A Reinforcement Learning Model for Understanding the Motivational Variation Underlying Avoidance

How do seemingly minor threats outweigh the obvious costs of avoidance? Pathological anxiety is characterized by avoidance of benign anxiety cues. It is important to understand the motivations underlying this seemingly irrational behavior.

This question is investigated from the perspective of Reinforcement learning (RL). RL is a computational framework for describing and understanding processes of learning and decision making. It has been widely used in many fields ranging from finance to machine learning. Recently it has also been applied to explain various processes in CBT. The RL system differentiates between two types of motivation. A reward signal is the immediate feedback that a learning agent receives when encountering a state. In contrast, the value function accounts for future implications and is responsible for planning. This distinction serves to analyze the decision making process leading to avoidance.

The current treatment describes three distinct motivational pathways: distress intolerance, core fears, and infinite repetition. Difficulty to tolerate distress has long been suggested as a mechanism of anxiety and as a target of exposure therapy. However, little has been written about the role of core fears and infinite repetition in motivating anxiety and avoidance.

Core fears are the ultimate underlying fears one has regarding their anxiety if they do not avoid or engage in safety behaviors. Typically, one accesses these fears via the downward arrow technique, repeatedly asking “and then what would happen?” until one arrives at an ultimate concern. Core fears can vary widely, for example, we have found that patients with OCD who engage in checking rituals do so for a variety of motivations including: fear of harm to oneself, fear of harming others, fear of losing precious objects, fear of losing control, and fear of social ostracization. Very little has been written about core fears despite their importance and common use clinically. Anxious individuals often report fear of having a feared outcome repeat “forever”. The mechanism underlying this process needs to be investigated more.

We show how these different processes naturally emerge from the structure of RL models. Furthermore, each of these motivational pathways calls for a different approach for designing therapy. We discuss the implications of these pathways for intervention and the understanding of the fear structure.