The Rescorla-Wagner theory does not predict contextual control of phasic responses in transswitching

Abstract

Theoretical predictions regarding differential phasic responding to the same phasic conditioned stimulus in two different tonic contexts, generated by a computer simulation of the Rescorla-Wagner theory of classical conditioning, were compared to empirical evidence of phasic switching in two studies of transswitching of the skin conductance response in humans. Significant phasic switching was found in both studies, but the Rescorla-Wagner theory substantially underestimated the obtained differences. It was shown that the theory predicts phasic differences that are more in accord with empirical data when the salience (alpha) of the contextual stimuli is assumed to be much greater than the salience (alpha) of the phasic conditioned stimulus. This modification in salience relationships, however, makes the Rescorla-Wagner theory more like Asratyan's theory of transswitching. Analysis of both tonic and phasic differentiation in experimental groups run with two different phasic stimuli AND two different tonic stimuli or with two phasic stimuli and only one tonic stimulus provided support for the conclusion that tonic response differentiation is a necessary precondition for obtaining phasic switching. It is suggested that the Rescorla-Wagner theory fails to account for phasic switching because it treats tonic and phasic stimuli as essentially equivalent, ignoring the fact that the tonic stimulus is present prior to the time that phasic stimuli occur and that responses to the tonic stimuli can occur during the period following tonic stimulus onset and preceding the administration of phasic stimuli.