Computational and Human Learning Models of Generalized Unsafety
Huskey, Alisa Mae
MetadataShow full item record
The Generalized Unsafety Theory of Stress proposes that physiological markers of generalized stress impair learning of safe cues in stressful environments. Based on this model, chronic problems inhibiting physiological arousal lead to a heightened perception of threat, which involves experiencing anxiety symptoms without any obvious precipitating stressful or traumatic event. This investigation aims to determine the impact of stressor- versus context-related emotional learning on generalized unsafety, using a Pavlovian threat-conditioning paradigm. The difference in learning threatening cues ([CS+] paired with an aversive stimulus) compared to safety cues ([CS-] not paired with an aversive stimulus) was used as a proxy measure of generalized unsafety, as conceptualized by the GUTS model. This difference is expected to be moderated by individual differences in tonic cardiac regulation (i.e. heart rate variability). Lastly, a temporal-differences learning model was used to predict skin-conductance learning during stressor, stressor context and general contexts to determine which best predicts Pavlovian learning. TD learning is expected to better predict skin-conductance in individuals with higher fear inhibition in comparison to those with low fear inhibition.
General Audience Abstract
This study examined the claims of a theory about how human bodies respond to stress and what this tells us about how anxiety develops in and affects the mind and body. The theory is named the Generalized Unsafety Theory of Stress (GUTS) and two main hypotheses were tested in this study: 1) the theory suggests that a person's feeling of safety is affected by the variation in their heart rate at rest, and 2) that a person's feeling of safety could be observed most accurately by their body's defense responses when they are experiencing a threatening situation that is objectively safe. Individuals experiencing anxiety often report being aware that they are safe, yet their heart rate remains elevated and palms remain sweaty. Most studies that have examined the body's defense response have focused almost solely on reactions to a threat by looking at the reactions of one or more organs that make up the body's defense-response systems (e.g., heart). Results of this study confirmed the unique GUTS perspective. Specifically, the heart rate's variation at rest affects the defense response (sweaty hands) during threatening and objectively safe contexts, which in turn, predicts a person's feeling of safety. These results confirm that there are measurable biological constraints that change the way people learn about and react to their environments, which is very important for understanding the development and maintenance of anxiety physiology and behavior. The way a person learns to associate emotional responses to certain cues in their environment, particularly threat and safety cues, can be measured as defense responses in the body in response to a series of trials. Exploratory analyses examined human threat learning in comparison with mathematically-generated learning in order to better model the processes whereby anxiety develops based on learning of threat and safety cues.
- Doctoral Dissertations