Individual differences in fear learning: Specificity to trait-anxiety beyond other measures of negative affect, and mediation via amygdala activation
Negative emotionality is a well-established and stable risk factor for affective disorders. Individual differences in negative emotionality have been linked to associative learning processes which can be captured experimentally in fear conditioning paradigms. Literature suffers from underpowered samples, suboptimal methods, and an isolated focus on single questionnaires and single outcome measures. Here, we apply multivariate and dimensional approaches for three commonly investigated questionnaires in the field (STAI-T, NEO-FFI Neuroticism, Intolerance of Uncertainty Scale) across multiple analysis units (ratings, skin conductance, startle, BOLD-fMRI) during fear acquisition-training in two large samples (NStudy1=356; NStudy2=113). We investigate whether the specific or shared variance of these questionnaires is linked with CS-discrimination in specific outcome measures (Study 1). We identify a significant negative association between STAI-T and CS-discrimination in SCRs and between Intolerance of Uncertainty and CS-discrimination in startle responding. Yet, correlation coefficients for all questionnaire-outcome measure combinations did not differ significantly from each other. In Study 2 the STAI-T score was positively associated with CS-discrimination in a number of brain areas linked to conditioned fear (amygdala, putamen, thalamus), but not to SCRs or ratings. Importantly, we replicate potential sampling biases between fMRI and behavioral studies regarding anxiety levels. We discuss the implications of these results.