Prior extinction increases acquisition context specificity in human predictive learning

Concurrent evidence of extinction making acquisition context specific and ABA and ABC renewal effects in human predictive learning.

“…The EMACS effect was also reported when P was trained in a context different from the extinction context (Rosas & Callejas-Aguilera, 2006; Experiments 2 and 3), and when it was trained after extinction in a completely different task (Rosas & Callejas-Aguilera, 2006; Experiment 4). Evidence of this effect in human predictive learning has been additionally reported by Ogallar et al (2019; see also Rosas, García-Gutiérrez, & Callejas-Aguilera, 2006). In humans, the EMACS effect has been also reported in other so-called automatic evaluation tasks (Ferguson & Zayas, 2009), in which organisms process environmental stimuli (a) extremely rapidly, (b) without intention, and in some cases (c) without awareness of making the evaluation (see Gawronski et al, 2015).…”

“…Since consistent, predictive relationships between foods and outcomes become quickly apparent in training, participants may be implicitly led to believe that one potential agent of action in the task-the patron-is irrelevant, and thus that another obvious potential predictive source-the restaurant context-may be important. Indeed, context-switch effects in the absence of extinction have been reported using this task (see Ogallar et al, 2019). One might worry that results from this task would not generalize well to other situations.…”

“…In other words, the goal of Experiment 1 was to test whether a cue–outcome association becomes context-dependent when a different cue is conditioned and then extinguished in two different contexts. To pursue this goal, we chose the new version of the predictive learning situation developed by Rosas and Callejas-Aguilera (2006), in which contexts and food were represented by actual pictures, rather than by their name (Ogallar et al, 2019). Two groups of participants were first trained with cue X followed by the outcome (O) in context A.…”

“…Diarrhea was used as the outcome (O). Contexts A and B were the images of two fictitious restaurants counterbalanced across participants (see Ogallar et al, 2019 for a more detailed description).…”

“…In one of the most commonly used tasks, the Restaurant task (e.g., , participants learn that customers eating different foods (cues) in specific restaurants (contexts) may develop a stomach illness (outcome). In this task, the food-outcome association usually transfers quite well across contexts in the absence of interference or ambiguity (but see Ogallar et al, 2019). However, when a food-outcome association becomes ambiguous (e.g., after an extinction treatment), the transfer between contexts is altered for both the extinguished cue and the rest of the cues that did not receive the extinction treatment.…”