Dissociating expectancy of shock and changes in skin conductance: An investigation of the Perruchet effect using an electrodermal paradigm.

Abstract

Is human Pavlovian conditioning driven by a unitary, propositional system (as claimed by Mitchell, De Houwer, & Lovibond, 2009) or by dual systems; one under conscious control, symbolic in nature, and requiring effort to deploy, and the other utilizing associative processes and automatic in its operation (McLaren, Green, & Mackintosh, 1994)? Past research has suggested that for electrodermal conditioning to occur in humans, conscious awareness of the contingencies is necessary to produce conditioned responding (e.g., Hinchy, Lovibond, & Ter-Horst, 1995), as predicted by single process theories that attribute the conditioned response (CR) to conscious expectancy of the shock. In this article, the authors examined the Perruchet effect (Perruchet, 1985), using an electrodermal paradigm to determine whether there is any role for associative processes in human electrodermal conditioning. The authors attempted to replicate the basic effect, whereby expectancy of an unconditioned stimulus (US) increases over a run of nonreinforced trials while the CR to the conditional stimulus (CS) declines, and the complementary pattern in which expectancy decreases over a run of reinforced trials while the CR to the CS grows in strength. In line with these patterns, the change in skin conductance response (our CR) as a function of US run length was found to follow a linear trend opposite to that of conscious expectancy of shock with respect to US run length. This dissociation supports a dual-processing system account of human Pavlovian conditioning, with conscious, controlled processes governing expectancy (and subject to the gambler's fallacy), whereas automatic, associative processes determine at least some of the strength of the CR to the CS.

Potentiation of the startle reflex is in line with contingency reversal instructions rather than the conditioning history

“…SCRs can be dissociated under certain conditions from cognitive contingency learning (Bechara et al, 1995;McAndrew, Jones, McLaren, & McLaren, 2012 Our conclusion runs counter the results of a number of studies that we mentioned in the introduction. We will discuss these studies in more detail here.…”

Dissociating conscious expectancies from automatic link formation in associative learning: A review on the so-called Perruchet effect.

“…There was no significant difference between negative and positive runs in McAndrew et al (2012), and the data reported by Williams and Prokasy (1977) went descriptively in the wrong direction (note, however, that the procedure and notably the reinforcement rate, were different). The only result clearly in line with the Perruchet Length and Type of the Preceding Run Figure 5.…”

“…In most studies, in keeping with Perruchet (1985), participants were instructed to rate during the intertrial interval (i.e., before the onset of E l) their expectancy that E2 will occur after the next E l (and not at the present time). Dismissing this measure amounts to considering that participants' expectancy of E2 after the next Elwouid change radically before and after the actual onset of E l. This hypothesis is not consistent with the fact that a between experiments comparison shows that ratings made before (e.g., Clark et al, 2001;Destrebecqz et al, 2010;Livesey & Costa, 2014;Perruchet, 1985;Weidemann et al, 2012) and after (McAndrew et al, 2012;Per ruchet et al, 2006) the onset of E l exhibit the same pattern. Moreover, at a more speculative level, the psychological processes that would motivate such a change in expectancy remain to be discovered.…”

Fear expression and return of fear following threat instruction with or without direct contingency experience

“…In the current study, however, CS-US pairings did not have the same effects on US expectancy and SCRs, which argues against the idea that both measures being affected by one common underlying (cognitive) factor. Dissociations between both measures have been reported before (e.g., Bechara et al, 1995;McAndrew, Jones, McLaren, & McLaren, 2012) while cognitive and emotional components of fear conditioning sometimes converge (e.g., Costa, Bradley, & Lang, 2014;Dawson, Rissling, Schell, & Wilcox, 2007), questioning the classification into cognitive versus emotional measures. Thus, carefully designed experiments employing a multimodal approach will be invaluable to further refine and develop models of associative learning.…”