Date of Degree


Document Type


Degree Name





Tracy A. Dennis-Tiwary

Committee Members

Evelyn Behar

Joel Erblich

Regina Miranda

Ekaterina Likhtik

Subject Categories

Behavior and Behavior Mechanisms | Clinical Psychology | Mental Disorders | Physiology | Psychological Phenomena and Processes


Anxiety disorders, Safety learning, Fear generalization, Attention bias


Anxiety disorders are the most prevalent mental health condition and affect one out of nine individuals around the world. Fear generalization is a neurocognitive mechanism thought to maintain and exacerbate anxiety, and thus is an important target of therapeutic interventions. Yet, intervention research and practice place relatively little emphasis on its importance. Given that a significant proportion of individuals do not respond to extant treatments of anxiety disorders, a strengthened focus on fear generalization may inform the development and personalization of new treatment approaches. Recent notions have linked fear generalization to failures in distinguishing between signals that predict the occurrence of an aversive event (threat cues) and those predicting the non-occurrence of an aversive event (safety cues). As such, a careful investigation of learning safety cues (i.e., safety learning) may advance our understanding of anxiety and fear generalization. The goal of the current dissertation tested the hypothesis that successful safety learning will reduce fear generalization. A secondary goal was to test whether individual differences in attention bias moderate the impact of safety learning on fear generalization. Attention bias, defined as selective and exaggerated attention toward threatening information and stimuli, varies across individuals and has been theorized to interact with learning processes to influence the development and maintenance of anxiety. This study examined a sample including individuals aged 18 to 45 years old (M = 24.05, SD = 6.07) who reported at least low to moderate levels of anxiety symptoms. Participants were randomly assigned to one of the following four groups: fear learning, safety learning without enhanced cue saliency, or one of the two additional safety learning groups with enhanced cue saliency. During a 2-day study period, all participants completed a fear/safety learning task, a fear generalization task, and a dot-probe task, with their subjective risk ratings, behavioral responses, skin conductance response (SCR), and event-related potentials (ERP) recorded simultaneously. Fear generalization was measured with risk ratings, SCR, and ERP. Attention bias was assessed with ERP and trial-level bias scores (TLBS) calculated using reaction times. Analyses revealed no sufficient evidence to support the predicted effect of safety learning procedures on fear generalization. Yet, individual differences were documented. Among participants who went through the safety learning with enhanced cue saliency, those who had attention bias towards threat exhibited reduced physiological responses to generalized and safety cues, suggesting reduced fear generalization. In sum, this dissertation provided initial evidence for the interplay between cognitive and learning processes in anxiety. Findings also demonstrated the importance of studying the impact of safety learning on fear generalization in the context of individual differences in disrupted processing of threat.