Marcelo Malbec

The aim of this dissertation was to investigate the influence of intolerance of uncertainty (IU) in shaping cognitive and affective responses across different experimental contexts. IU is considered a transdiagnostic construct for internalizing disorders, thus understanding its influence can improve our understanding of its impact on anxiety and depression. In Chapters 2, 3, and 4, I present results from investigations on three cognitive processes: performance monitoring, fear conditioning, and reward processing. Each study used a combination of behavioral and physiological measures to assess how individual differences in IU and its subfactors (i.e., prospective and inhibitory IU), relate to affective and cognitive processes relevant for internalizing psychopathology. In Chapter 5, we adopted a data-driven approach to identify distinct neurocognitive profiles underlying individual differences in the cognitive processes previously described. Leveraging the comprehensive assessment made possible by the use of the same sample of participants across all tasks, this approach allowed us to uncover latent patterns that could explain variability in intolerance of uncertainty and related mental disorders.

Performance monitoring and IU: Chapter 2 focused on performance monitoring using a modified Flanker task. Neural and behavioral indices of performance monitoring showed the expected patterns with more negative amplitudes for errors than correct trials for the ERN and the opposite pattern for the Pe, as well as a significant increase in reaction times after errors compared to correct responses (i.e., post-error slowing). These findings confirmed the sensitivity of the task to detect performance monitoring responses. Regarding IU and its subfactors, they were not associated with the ERN or Pe, neither predict post-error adjustments. However, a small but significant association was observed between total IU and the correct-response negativity (CRN), with higher IU predicting more negative CRN amplitudes. This effect remained after controlling for trait anxiety, suggesting a degree of specificity to IU. Exploratory moderation analyses also found no significant interactions between IU and depression, trait anxiety, or worry in predicting ERN amplitude. Together, the findings suggest a limited role for IU in modulating performance monitoring, with one exception emerging for correct-response processing.

Fear conditioning and IU: Chapter 3 focused on both fear acquisition and extinction using a differential conditioning paradigm with immediate extinction. During fear acquisition training, successful fear conditioning was indicated by increased SCRs and ssVEP amplitudes as well as increased subjective ratings (valence, arousal, and threat expectancy) towards the threat cue (CS+). During acquisition, LPP amplitudes were larger for the safe cue (CS−) than for the threat cue (CS+) in the first half of acquisition, a pattern that then disappeared, showing no differences between cues during the second half. No significant associations were found between IU (total or subfactors) and fear learning at this stage, across any of the outcome measures. These findings are consistent with prior studies reporting no effects of IU during acquisition and may reflect the limited influence of IU under high-threat, high-contingency conditions. During extinction, physiological and self-report responses showed partial extinction of conditioned responses. SCRs no longer differentiated between CS+ and CS− at late extinction, indicating successful extinction learning. In contrast, LPPs, ssVEPs, and valence/arousal ratings continued to show larger responses to the CS+, suggesting persistence in subjective and cortical indices of threat discrimination. However, IU did not significantly modulate extinction dynamics in this sample, contrasting with earlier studies that found some evidence of delayed or impaired extinction in high IU individuals. Overall, these findings suggest that IU plays a limited role in differential fear learning and extinction in the context of this study, which used a relatively high reinforcement rate and immediate, uninstructed extinction.

Reward Processing and IU: Chapter 4 focused on reward processing under uncertainty using a modified version of the Doors Task. Participants received monetary feedback (reward or no-reward) in this task after choosing among options with varying levels of reward outcome probability (low, high, and ambiguous risk). Feedback-related ERPs showed robust modulation by feedback type and risk level, with more positive amplitudes following reward compared to no-reward outcomes, and higher amplitudes under high-risk conditions. IU total scores were not significantly associated with electrocortical responses. However, the IU subfactors showed independent and opposing effects: prospective IU was associated with increased ERP amplitudes following both reward and no-reward feedback, while inhibitory IU predicted reduced amplitudes across both conditions. These effects remained after controlling for depressive symptoms, trait anxiety, and worry. Previous studies have also reported opposing effects of IU subfactors on reward processing markers. However, those effects were typically observed in difference scores between feedback types, whereas our findings reflect an impact of IU subfactors on overall neural responses across feedback conditions. Additionally, depressive symptoms were associated with blunted amplitudes specifically following no-reward outcomes. The findings highlight distinct associations of IU subfactors with feedback-related neural responses, independent of overall internalizing symptomatology.

Neurocognitive Profiles Across Tasks: Chapter 5 used latent profile analysis (LPA) to combine the data across the three cognitive processes to identify subgroups of participants with similar patterns of responding. The analysis yielded three distinct profiles based on behavioral and neurophysiological indicators of performance monitoring, fear and extinction learning, and reward processing. Profile 1, High Threat and Performance Reactivity with Low Outcome Reactivity, showed larger ERN/CRN and Pe, higher ssVEPs to threat and safety cues, and reduced feedback-locked ERPs to outcomes. Profile 2, High Outcome Reactivity, showed strong feedback-locked ERPs with otherwise typical responses in threat and cognitive control. Profile 3, Average Reactivity, was close to the sample mean across most measures, with a slight reduction in ssVEPs through fear conditioning. These profiles, however, were not different on measures of IU, including both subfactors, nor depressive symptoms or worry. State anxiety was modestly higher in the more reactive profile, but this difference should be interpreted with caution. Overall, the profiles capture meaningful variation in neurocognitive responding, while links with trait questionnaires were limited.