Excessive and Abstract Spread of Fear
Improving the Regulation of Fear
Real-World Dynamics of Anxiety and Trauma
Merging Experimental and Quant Methods
Examples of recent research questions:
How do patterns of generalized fear and avoidance, and their covariation, relate to adaptive and maladaptive personality structures?
What is the state of the evidence for excessive fear generalization as a pathological correlate of anxiety-related disorders?
Through which neural mechanisms do we integrate past memories and new fearful experiences, and do these mechanisms predict psychopathology?
Can we improve on standard extinction methods using rewarded extinction (i.e., aversive-to-appetitive counterconditioning) to improve safety learning and decrease threat reinstatement in PTSD and related psychopathology?
Do lab-based neural threat indices predict latent psychopathology dimensions and their interactions with anxiety and trauma symptoms measured multiple times per day (i.e., real-world dynamics)?
How do higher-order (i.e., indirect or abstract) forms of fear generalization and its extinction relate to PTSD and OCD symptoms?
Example of acquisition, extinction, and generalization phases and example data; adapted from Cooper et al., 2022, Neuropsychopharmacology
Most of our work involves the building and implementation of experimental tasks that probe psychopathologically-relevant processes. In particular, we make use of classical and instrumental conditioning paradigms to study emotions and behaviors. Given the lab foci, most of the time we are using these techniques to study fear and avoidance. In the acquisition portion of thesee tasks, we pair an inherently aversive stimulus (e.g., mild electric stimulation, US) with a neutral stimulus (e.g., a generic picture or sound, CS) until the neutral stimulus is no longer neutral - it now provokes a negative emotional or behavioral response without any help from the original aversive stimulus (now termed CS+). These responses, formally known as conditioned responses, are what we measure and what are thought to underlie the distress and impairment seen in many forms of anxiety. We can then either remove the US to promote safety learning (extinction) or introduce CSs that are similar to the CS+ (generalization). These techniques are rough analogues of how we develop fear (acquisition), fear maladaptively spreads (generalization), and is then maintained or regulated over time (extinction).
There are many ways to measure conditioned responses or other indices of learning, memory, or related cognitive variables. In the lab, we use physiological (e.g., startle blink, sweat response, heart rate) measures, as well as subjective ratings of threat, contingency expectancy, and affect. In the past, Dr. Cooper has also used eye-tracking and pupillometry to assess attentional and arousal-based conditioned responses, as well as attentional and perceptual processes. Currently, much of the lab is devoted to studies that use functional magnetic resonance imaging (fMRI), a noninvasive neuroimaging technique we use to assess conditioned responding at the level of the brain and its constituent circuits. A K23 from NIMH currently funds work in this area, including the implementation of an fMRI study of category sensory preconditioning. This involves the use of a novel task developed in conjunction with the Dunsmoor lab, in which we employ multivariate pattern analysis to decode latent brain states and subsequently identify these decoded patterns during subsequent parts of our tasks.
Schematic of an fMRI optimized “category preconditioning” paradigm that uses MVPA decoding techniques to derive indices of neural similarity and integration, based on Cooper et al., 2024, Current Biology.
Graphical reimagining of a table showing predictive relations between latent maladaptive personality factors (PID-5) and different types of OCD symptoms (Checking, Hoarding, Neutralizing, Obsessing, Ordering, Washing) across two samples; adapted from Cooper et al., 2022, Assessment
Although experimental techniques are at the core of our work, it is in combination with individual differences sciences that allows us to provide compelling answers to scientific questions regarding the nature, development, and treatment of psychopathology. In particular, we use dimensional models of psychopathology—which treat mental illness as just the extreme or less common part of a continuum of human behavior and internal processes, as opposed to qualitatively distinct categories—to understand how pathological processes differ across and within people. Part of this work involves understanding how psychopathology is organized and structured in the first place. Recent work has involved using latent variable modeling (e.g., EFA, mixture modeling, SEM/eSEM/mSEM) to more precisely capture the underlying psychological dimensions that span a variety of symptoms or conditions, specifically those commonly defined as anxiety-related disorders. This work aligns with the Hierarchical Taxonomy of Psychopathology (HiTOP), an effort to improve our models of psychopathology, and multiple lab projects are conducted in collaboration with HiTOP.
A replication crisis, re-evaluation of applied quantitative techniques, and a rapid increase in readily available computational power, among other things, have exponentially increased emphasis and interest on ensuring we use the most robust and generalizable methods whenever possible. This is reflected across most of our work, and in some cases, it is explicitly the topic or focus of specific project. Work in this area includes psychometric analyses of threat conditioning (e.g., test-retest reliability), expanding the repertoire and rigor of (generalized) mixed-effects regression employed in experimental work, advocating for multivariate analyses of fMRI data, and recommendations for best design and analytic practices in neuroscience-informed psychopathology research. Part of improving our science is also systematically and quantitatively evaluating our methods and results, commonly referred to as meta-science (i.e., the science of science). Meta-analysis (statistical analysis of many previous results) is an important tool for us, and we are increasing our participation in other forms of quantitative and qualitative meta techniques (e.g., consortium-wide mega analysis, meta-reviews).
Best practices for measurement, design, and analysis using modern individual difference methods, from a HiTOP review and primer on precision phenotyping in psychiatry and psychopathology, Tiego et al., 2023.