UCLA

Goal: The goal of this dissertation is to investigate threat and fear responses in humans in order to understand both the multifaceted nature of these responses and how they can go awry in fear and anxiety related disorders. By understanding the mechanisms behind fear and anxiety disorders, effective treatments can be designed that minimize distressing or panic-inducing experiences that too often lead to attrition from the clinic. These issues are investigated across three studies utilizing behavioral tasks and neuroimaging via functional magnetic resonance imaging (fMRI) with the following aims:

Aim 1. Establishing efficacy of decoded neurofeedback as an intervention for specific phobia in a clinical cohort. There is an unmet need for non-distressing treatments for fear-related disorders like specific phobia where exposure therapy is the gold standard. While the effectiveness of exposure therapy should not be diminished, it is an inherently distressing experience leading to high rates of attrition and a need for alternative options for those who can not tolerate the experience. In the first study of this dissertation, I tested multi-voxel neuro-reinforcement as an intervention for specific phobia in a randomized double-blind placebo-controlled clinical trial. I found evidence for reduced amygdala activation to phobia as well as less attentional capture by the target phobia in an affective stroop task post-treatment.

Aim 2. Classifying the subjective awareness of threat from multi-voxel patterns. Most studies of threat processing have relied on direct contrasts between conditioned stimulus types as a proxy for threat detection. However, such contrasts fail to take individual differences in threat learning success and generalization into consideration. In the second study of this dissertation, I use machine-learning techniques to classify the subjective awareness of threat from whole-brain multi-voxel patterns. Additionally, I classify threat awareness iteratively within brain regions to identify which brain regions are most critical for threat awareness. These analyses reveal a distributed brain response to threat that is organized hierarchically along the visual stream. Results are also characterized in terms of self-reported participant symptomatology.

Aim 3. Investigating brain networks involved in acquisition, extinction, and recall of learned threat. While many brain regions have been implicated in the formation of threat memories, the whole-brain dynamics of learned threat in humans are still poorly understood. In the third study of this dissertation, I applied group independent component analysis to whole-brain fMRI data. I identified brain networks involved in the acquisition, extinction, and recall of learned threat in a Pavlovian threat conditioning paradigm. This revealed stable networks across task phases that responded in opposing fashions across the same timescale. Results from this study highlight the multitude of parallel processes and networks that are engaged in human threat detection and learning.