The control of spontaneous recovery of anxiety in imaginal desensitization

Abstract

One of the practical difficulties encountered in imaginal desensitization is that an item that has been completely desensitized in one session can show some spontaneous recovery of anxiety by the next (Agras, 1965; Rachman, 1966). No doubt the extent to which this occurs is largely related to the subject's phobic experiences between sessions. But it is also probably affected by the way desensitization is carried out. For example the amount of anxiety decrement that survives spontaneous recovery is greater if 30 rather than 5-sec item presentations are used. In contrast it does not seem to be affected by interstimulus inverval length (Watts, 1973). it also seems possible that the subject's arousal level during desensitization may affect the amount of spontaneous recovery. Teasdale (1971) found that relaxation produced more spontaneous recovery in auditory habituation, and the same may apply in desensitization. It is obviously of considerable importance to understand fully the factors that control such spontaneous recovery. The present research reveals another relevant procedural variable. The starting point was the previous finding (Watts, 1971) that item presentation lengths is important in this connection. One possible explanation of this is that longer presentations provide greater opportunity for stimulus learning or familiarization of some kind. The extent to which desensitization is accompanied by such familiarization may influence how much spontaneous recovery of anxiety occurs subsequently. A parallel can be found in the effects of stimulus differentiation on verbal learning (Saltz, 1971). In Saltz's revision of stimulus differentiation theory, a well differentiated stimulus is less likely to attract an irrelevant response that will interfere with the correct one after learning. In the same way a well differentiated desensitization item may be less likely to re-attract anxiety after the treatment session. One way of investigating this hypothesis is to employ a procedural variable that directly affects the opportunity for stimulus differentiation in desensitization. This can probably be done through the way an item is presented. If it fully described each time the subject has to imagine it, stimulus differentiation should be greater than if he is merely told to imagine ‘the same’ items again. Spontaneous recovery of anxiety should be reduced as a consequence. Watts (1971) also found that longer item presentations resulted in faster desensitization of ‘high’ items. This may also be due to better stimulus learning, though the rationale for this is less clear. The effects of item re-description were therefore examined both on the speed at which high items desensitized and on the amount of anxiety reduction that survived spontaneous recovery. It seemed possible that the benefit of item redescription might be greater with longer presentations. To examine this and also to replicate the original findings, stimulus length was used as a second independent variable. Different values (45 and 15 sec) were used from those used previously (30 and 5 sec). The two independent variables were combined in a (2 × 2) factorial design and the resulting conditions rotated across desensitization sessions in a series of five individual-centred experiments.