Contemporary cognitive models of classical conditioning provide a framework for understanding people’s fears and phobias. According to these models, classical conditioning is the acquisition of associations between representations of stimuli or events. For an explanation of classical conditioning, see a number of previously summarized articles.

Recently a distinction has been established between two different forms of learning in classical conditioning: expectancy learning and merely referential learning. Expectancy learning arises when showing the CS activates the expectation of a real US event. Referential learning arises when showing the CS activates the representation of the US, without the expectation of the US also arising. So the CS ensures that someone thinks of the US, but doesn’t expect the US.

Evaluative conditioning (EC) is a type of referential learning: the observation that the presentation of neutral stimuli with (dis)liked stimuli changes the valence of the stimuli in a (negative) positive direction. Evaluative learning resembles expectancy-learning in that it is sensitive to US revaluation. So the shift in valence of the CS seems to be based on the association between Cs and US representations. Research into evaluative learning and expectancy learning differs in two areas:

1. The type of learning: evaluative conditioning studies often used rather unobtrusive stimuli like flavours, or pictures of human faces, paintings, statues etc. that were rated as positive or negative by the individual participant. Whereas expectancy-learning have utilized biologically significant stimuli like food or shock, that are unconditional stimuli in a strict sense: they elicit a similar and innate/unlearned positive or negative response in all subjects.
2. Response systems are different: evaluative investigations focus primarily on the verbal ratings of CS valence and expectancy investigations primarily on the nonverbal motor/autonomous responses such as the skin conductance response.

In addition to verbal assessments, it’s important to investigate nonverbal and physiological responses. A good way is the performance on affective priming tasks. Positive or negative stimuli are presented, which the participant must give a value as quick as possible to (positive or negative). Each stimulus is preceded by a prime stimulus, which is negative, positive, or neutral, and must be ignored. The valence of the prime stimulus appears to influence the speed at which the stimuli are assessed. This is based on the automatic processing of the prime stimulus.

Current research is looking into whether evidence can be found for both forms of learning within the same conditioning procedure. This conditioning procedure includes images of human faces as CS+ and CS- and an electrocutaneous stimulus as US. The valence was investigated through the affective priming procedure and verbal assessments. Expectancy learning has been examined by assessments of US-expectancy. Arousal and fear ratings were also assessed for exploratory reasons.

In addition, this study aims to analyze whether evaluative learning as demonstrated in aversive conditioning is comparable to evaluative learning with non-aversive US. In addition to the CS+ and the CS-, therefore, CS-positive and CS-negative were included, positive and negative adjectives, respectively. Finally, this study looks at the extent to which specific instructions for imagining can enhance learning. Previous research has shown that mental imagery can potentiate the effects of acquisition.

Method

32 students (11 men, 21 women) were examined.

Experiment consisted of three phases:

1. Stimulus selection: participants assessed 60 photos of faces. The intensity of the US was also measured in this phase by means of electrodes.
2. Acquisition: CS+, CS-, CS-positive, and CS-negative were shown to the participants. A corresponding sound was played during presentation of CS-positive and -negative.
3. Affective priming: the affective priming task was offered.

Results & Discussion

After the acquisition, the participants indicated that the "electrocutaneous US" was intense and unpleasant and that they were shocked. During the acquisition, the CS + had become a valid predictor of the US. Not only did the participants find the US annoying and they were aware of the contingency between CS + and US / CS- and not a US, expectation learning had also emerged. In addition, there have been shifts in the evaluation of the CS + and CS-. After acquisition, the participants gave the US with positive adjectives a positive rating and the negative US a negative rating. No expectation learning is involved in the conditioning of the CS positive and the CS negative. Moreover, no significant difference was found in the degree of evaluative learning between the two types of conditioning procedures.

To increase the level of learning, half of the participants were asked to imagine something about the relationship between the CS and the specific US. It turned out that the two groups did not differ in the degree to which they liked or disliked the different types of US. There was also no significant effect on the level of evaluative learning. The degree of arousal was influenced by the conditioning procedures: the CS + was found to be more arousing than the CS-. The CS + also generated more fear than the CS-. The CS negative generated more fear than the CS positive.

In conclusion, the experience of contingent presentation of a neutral face (CS +) and an aversive electrical stimulus (US) has changed the experience of the CS + in two ways:

• The CS + became a predictor for the US.
• The CS + itself became more negative.

Thus, evaluative learning and expectation learning can occur together. Moreover, the results of the affective priming procedure prove that affective priming can occur with aversive conditioned stimuli. An alternative to this affective priming procedure is the modulation of an acoustic shock reflex: a defensive reflex due to an unexpected sound. This varies with the valence that one gives to a picture. This modulation can be a good way to measure valence after acquisition, but it cannot measure the effects of extinction. A substantial fright reflex is only found with pictures that cause a lot of excitement, and they usually disappear after extinction.

In addition to the fact that expectation learning and evaluative learning can occur together, a conclusion can also be drawn from the results found. This is that the amount of evaluative learning resulting from an aversive conditioning procedure has no interaction with the type of conditioning preparation. Thus, evaluative learning can also be generalized to conditioning procedures with more biological US stimuli. This shows that clinical fears based on the experience of contingency between the neutral stimulus and negative stimulus cause not only that the CS becomes a predictor of a negative experience, but also that the CS is viewed more negatively. A combination of exposure and counter conditioning would be a good way to prevent relapse of patients.

A number of things can be interesting for follow-up research. Firstly, to investigate whether the conditioning effects found are also found in other types of stimuli. Secondly, to investigate whether the same evaluative shifts are found if the CS is subsequently masked during acquisition. Third, to investigate the precise nature of the conditioned response, which remains after extinction.