(PS9-A16) The Utility of a De novo Fear Conditioning Paradigm as a Predictor of Exposure-based Therapy Outcomes: Null Results Add to an Already Inconclusive Literature

Exposure-based therapy (EBT) is an efficacious treatment for anxiety- and fear-based disorders; nevertheless, many patients fail to respond adequately to treatment. Laboratory-based models of fear learning (de novo fear conditioning) are thought to be an important translational research tool for elucidating the mechanisms underlying EBT and for studying the development and maintenance of fear-related disorders (Carpenter et al., 2019). However, relatively few studies have examined whether de novo fear conditioning paradigm outcomes are indeed associated with EBT outcomes. We and others have identified only 7 studies investigating these associations; results to date show inconsistencies in indices of fear conditioning examined (across acquisition, extinction, and recall phases) as well as poor replication between studies in the task of predicting EBT outcomes from de novo conditioning data (Scheveneels et al., 2021). The current study was designed to further investigate this issue by using a comprehensive set of threat/fear conditioning predictors and focusing on a fuller protocol of EBT for adults with social anxiety disorder.
Our outpatient sample was from a double-blind, randomized clinical trial (Hofmann et al., 2019). Patients completed a 3-day de novo fear conditioning and extinction paradigm and, later, a 5-session group EBT. Fear conditioning outcomes were indexed with skin conductance response (SCR) and expectancy ratings. Fifty-one individuals who showed valid SCR in the acquisition phase of the fear conditioning paradigm were included in analyses of treatment completers.

Multiple regression analyses were used to examine the relationships between fear conditioning predictors (e.g., average differential SCR during extinction) and the clinical outcome of LSAS score at the 1-week follow-up. Despite evidence of adequate acquisition and extinction for the sample, results of the regression analyses revealed that no SCR or expectancy variables for the acquisition, extinction, and recall stages of the conditioning paradigm offered significant prediction of clinical outcomes from EBT (all p-values > .08).

Although the literature is limited by smaller scale studies, our findings help confirm that there is no consistent finding across any two studies among the 8 studies (including the current study) that have investigated outcomes from de novo fear conditioning as predictors of EBT outcomes. Accordingly, it is time to more assertively question whether the individual differences observed for acquisition, extinction, and recall of fear in de novo fear conditioning paradigms provide a useful analogue for the type of learning encountered in EBT for fear-related disorders. Replication of measures across studies and the use of larger sample sizes is needed to help confirm these preliminary concerns. In addition, limitations on power be discussed further.