Dissertations, Theses, and Capstone Projects

Date of Degree

6-2026

Document Type

Master's Thesis

Degree Name

Master of Science

Program

Cognitive Neuroscience

Advisor

David Johnson

Subject Categories

Cognitive Neuroscience | Cognitive Psychology | Cognitive Science

Keywords

Fear, Learning, Fear conditioning, Uncertainty, Anxiety, Individual differences

Abstract

Trait measures of negative emotionality have sometimes been shown to play a role in shaping threat learning, but findings are mixed. For example, Johnson et al (2023) showed that the inhibitory and prospective subscales of the Intolerance for Uncertainty Scale (IUS) were associated with diminished and enhanced threat-safety discrimination, respectively, during the acquisition phase of a Pavlovian fear conditioning task. This finding was subsequently replicated in an independent sample (Kinney-Petrucha, 2024, unpublished). However, these findings and other trait-level fear acquisition effects have not been consistently observed in the broader fear learning literature. Variation in the complexity of or uncertainty inherent to the experimental design have been proposed as explanations for these disparate findings. It has been suggested that increasing the complexity of the design (e.g., increasing the number of cues and variability between them) constitutes a “weakening” of threat acquisition, which could cause individual differences to be more likely to surface. However, this idea has yet to be widely tested. We hypothesized that using a simpler two-cue design would attenuate the individual difference effects observed in Johnson et al (2023) and Kinney-Petrucha (2024), which utilized a more complex three-cue design. Furthermore, as experimental complexity and uncertainty are partly overlapping constructs, we also varied the uncertainty of the threat cue, positing that even when utilizing a simple two-cue design, the IU subscale effects observed previously might surface if the threat cue is highly uncertain. Thus, we used a one-day, Pavlovian threat learning design with two cues (one CS+ and one CS-) and randomly assigned participants (n = 162) to one of four conditions: no, low, moderate or high uncertainty for the threat cue. All participants completed the Intolerance of Uncertainty Scale (IUS) and the trait portion of the State Trait Anxiety Index (STAI-T). While there was diminished threat-safety discrimination for the certain vs uncertain threat cues, we did not observe the emergence of any individual difference effects related to the level of uncertainty of the threat cue nor did we replicate the aforementioned trait levels effects in the overall sample. However, we unexpectedly found a statistically significant (but small) reversed effect for the Inhibitory Intolerance of Uncertainty subscale (I-IU), such that higher scores were associated with enhanced threat–safety discrimination in the simplified task context. This suggests that the influence of the IU subscales on threat-safety discrimination learning may depend on features of the learning environment, such that high inhibitory IU can diminish learning in complex environments, whereas it can enhance learning in simpler ones, perhaps due to the more manageable environment facilitating increased vigilance or heightened attention to motivationally relevant stimuli. This finding yields some important considerations. For one, fear/threat researchers interested in studying individual differences in threat acquisition should carefully consider the complexity of their experimental designs. While more simplistic designs tend to produce better discrimination learning, they could attenuate or even reverse individual difference effects that would otherwise surface when using more complex designs. Second, variation in experimental design complexity may partly account for the heterogenous individual difference findings reported in the broader fear learning literature.  Future meta-analyses may benefit from explicitly modeling design parameters related to the complexity of the learning environment when synthesizing results across studies. Finally, these findings may extend beyond fear/threat learning to other cognitive domains exploring individual difference effects. Although this finding is promising and the study was well powered, replication will be essential to strengthen the empirical support for these conclusions.

Download

Share

COinS