Methodological strengths of the current study are at least 3-fold. An important merit of this study is that it is the first to demonstrate impaired fear inhibition in healthy individuals scoring relatively high on TA in a fear conditioning paradigm using dynamic proprioceptive stimuli (eg, joystick arm movements) instead of more static exteroceptive (eg, auditory or visual) stimuli. Basically, we were able to extend the findings on deficient safety learning to the field of pain, which
might be particularly relevant to spreading of painrelated fear in chronic pain. Second, this study focuses on resilience and not merely on vulnerability during exposure treatment. This is the first experimental study on (extinction of) fear of movement-related pain that demonstrates that a positive trait can counteract the detrimental effect of a vulnerable trait, that is, TA.
Finally, we used multilevel regression analyses including a random intercept parameter to capture individual differences. Such models treat the trait variables as continuous variables, which maximizes theuse of information, as compared with, for example, a median split model, leading to a favorable model fit and high portion of explained variance (R2 = .77).
Further, we specified a model that incorporates both TA and trait PA, and we corrected for the level of learning at the end of acquisition