Evaluating the learning of stimulus-control associations through incidental memory of reinforcement events.

Abstract

Cognitive control describes the ability to use internal goals to strategically guide how we process and respond to our environment. Changes in the environment lead to adaptation in control strategies. This type of control-learning can be observed in performance adjustments in response to varying proportions of easy to hard trials over blocks of trials on classic control tasks. Known as the list-wide proportion congruent (LWPC) effect, increased difficulty is met with enhanced attentional control. Recent research has shown that motivational manipulations may enhance the LWPC effect, but the underlying mechanisms are not yet understood. We manipulated Stroop proportion congruency over blocks of trials and after each trial, provided participants with either performance-contingent feedback ("correct/incorrect") or non-contingent feedback ("response logged") above trial-unique, task-irrelevant images (reinforcement events). The LWPC task was followed by a surprise recognition memory task, which allowed us to test whether attention to feedback (incidental memory for the images) varies as a function of proportion congruency, time, performance-contingency, and individual differences. We replicated a robust LWPC effect in a large sample (N = 402), but observed no differences in behavior between feedback groups. Importantly, the memory data revealed better encoding of feedback images from context-defining trials (e.g., congruent trials in a mostly congruent block), especially early in a new context, and in congruent conditions. Individual differences in reward and punishment sensitivity were not strongly associated with control-learning effects. These results suggest that statistical learning of contextual demand may have a larger impact on control-learning than individual differences in motivation.

Minimal impact of consolidation on learned switch-readiness.

“…Further, we replicated Monsell and Mizon’s (2006; Experiment 4) finding of between-participants LWPS effect in a more high-powered sample and using unbiased items when we analyzed the first four experimental blocks alone and treated listwide proportion switch context as a between-participants factor. This finding had also previously been documented for the listwide proportion congruent effect (Bejjani & Egner, in press; Spinelli et al, 2019) and in a hybrid free- and forced- choice setting (Fröber & Dreisbach, 2017, ).…”

“…This allowed us to probe whether the LWPS effect can be observed between participants. A between-participants LWPS effect has been reported in in one previous study that did not use a design with unbiased items (Monsell & Mizon, 2006, Experiment 4), and in a hybrid free- and forced-choice setting (Fröber & Dreisbach, 2017), as well as in studies of the proportion congruency effect (Bejjani & Egner, in press; Spinelli et al, 2019). We did this by analyzing only the first PS context (the first four blocks) of each session, that is, before participants experience a PS context shift, and treating PS context order as a between-participants factor.…”

“…Together, these findings suggest that the acquisition of stimulus-switch-readiness associations is stable within participants, and can be learned relatively quickly (cf. Bejjani & Egner, in press), but is not benefited by long-term memory consolidation (or “spacing” of learning). In the following, we discuss different possible reasons for why learned switch-readiness may not have benefited from the consolidation period in our study.…”

The many faces of learning-guided cognitive control.

“…An interesting question raised by these findings is whether statistical learning about the likelihood of control demand in general is sensitive to reinforcement or other motivational influences. Bejjani and Egner (2021) used incidental encoding of feedback events to probe how/whether reinforcement plays a role in control learning. The upshot was that people seem to engage in building a statistical prediction of trial types (e.g., expecting congruent trials in mostly congruent lists) and that confirmation of those predictions works as reinforcement, thus promoting control learning.…”

“…As is evident from these studies, control learning takes various forms, and shapes how people overcome conflict (Bejjani & Egner, 2021; Chen et al, 2021; Spinelli & Lupker, 2021; Suh & Bugg, 2021) and switch flexibly (Bejjani et al, 2021) based on the statistics of the environment. The study by Trach et al (2021) suggests that learning about likely demands can occur at several levels (possibly independently), including one that sits above the level of single task sets, specifically learning about sequences of tasks.…”