Abstract (eng)
In the European population about one-fifth of individuals experience fear of spiders and in many cases the resulting distress interferes with everyday life. The current state-of-the art treatment option, exposure therapy (ET), shows high drop-out rates, a large variation in practice and is thus suboptimal. Therefore, we aim to develop an efficient computer-assisted ET paradigm that involves adaptation of the stimuli (images) to the individuals’ fear states, which can be operationalized for example by physiological arousal, or subjective fear. The present study lays the groundwork for such a development. We exposed spider fearful individuals (N = 51) to a novel database comprising spider-related images, and assessed subjective fear and skin conductance response (SCR) for every stimulus. Treatment effects were examined by means of standardised pre- and post-questionnaires (Fear of Spiders Questionnaire; Spinnenangst-Screening, SAS), and by repeated collection of ratings on the variables fear, and disgust. Firstly, the results revealed habituation effects for SCR and for the SAS. Secondly, SCR and mean fear rating were identified as predictors of subjective fear by means of linear as well as non-linear machine learning models (regularised linear least absolute shrinkage and selection operator regression; ensemble of randomized trees analyses). Thirdly, high-fearful individuals reported significantly higher fear levels in response to the exposure than low-fearful individuals. Thus, it was confirmed that the novel database has a fear-inducing quality in spider-fearful individuals. In addition, it was shown that SCR patterns have a predictive value with regard to subjective fear but that further and possibly more precise predictors should be explored to eventually develop an adaptive ET paradigm. Most importantly, the study provided basic insights into the mechanisms underlying spider fear and guided the generation of objectives for follow-up research.