There are a number of different explanations of why systematic desensitization reduces anxiety (Spiegler & Guevremont, 1998). We will describe one of them using ourexample of Nancys public speaking phobia to illustrate. Relevant to this analysis is the concept of counterconditioning: a method of Pavlovian conditioning in which an undesired CS-CR relationship is weakened or eliminated by setting up a new CS-CR relationship in which the new CR is opposite to or incompatible with the old CR (Grant & Evans, 1994, p. 383).
(The following analysis requires a knowledge of classical conditioning; if you are not familiar with this concept, then consider first completing the classical conditioning turorial.)
Systematic desensitization, in addition to breaking the problematic CS-US contingency responsible for the negative emotional response, forms another potentially advantageous contingency between two stimuli.
Overall, the extinction of the problematic CS-US contingency (Item #'s 1-3) and the formation of a new CS-US contingency (Item #'s 4-6) are hypothesized to have the following effect: when speaking in class, Nancy feels relaxed rather than anxious.
In systematic desensitization, the negative emotional response is extinguished in the presence of gradually more intense conditioned stimuli by repeatedly presenting each stimulus without following it with the US that elicits the negative emotional response. Wolpe (1958) describes the theory behind the graded series of steps.
If a stimulus constellation made up of five equipotent elements A1A2A3A4A5 evokes 50 units of anxiety response in an organism, proportionately less anxiety will be evoked by constellations made up of fewer elements. Relaxation that is insufficient to counter 50 units of anxiety that A1A2A3A4A5 evokes may well be able to inhibit the 10 units evoked by A1 alone. Then if the anxiety evoked by A1 is repeatedly inhibited through being opposed by relaxation, its magnitude will drop, eventually to zero. In consequence, a presentation of A1A2 will now evoke only 10 units of anxiety, instead of 20, and this will similarly undergo conditioned inhibition when opposed by relaxation. Through further steps along these lines the whole combination of A1A2A3A4A5 will lose its power to arouse any anxiety. (p. 139)