Learning what to expect: context-specific control over intertrial priming effects in singleton search

Abstract

The present study explored the degree to which repetition effects in color pop-out search from trial n − 1 to trial n are subject to the attentional control settings of the observer. Intertrial priming effects were compared between two contexts that differed in terms of the utility of immediate prior experience for current performance; in one context, the target was likely to repeat, and in the other context, the target was likely to alternate from one trial to the next. Across two experiments, priming of pop-out (PoP) effects (Malkjovic & Nakayama; Memory & Cognition 22:657-672, 1994) were modulated in accord with the probability of target color repetition in a given trial context. Importantly, this modulation persisted when trial history preceding trial n − 1 was controlled for. Furthermore, this control over PoP seems not to derive from explicit strategies and is not an artifact of randomly occurring strings of same-target trials. We argue that priming effects in singleton search from trial n − 1 to trial n are subject to a form of implicit top-down control.

Comprehension priming as rational expectation for repetition: Evidence from syntactic processing

“…We have shown that such adaptation is sensitive to higher-order temporal contingencies between sentences, supporting a role for multiple levels of linguistic context in rapid implicit learning throughout discourses. Multilevel context is consistent with evidence from other domains, in which sensitivity to both the base rate of a stimulus as well as higher-order sequential patterning are jointly necessary to account for human behavior (Thomson et al, 2012; Wilder et al, 2009). …”

“…More experience led to increasing reliance on the repetition pattern rather than the identity of the just-prior stimulus, suggesting that participants were learning and adapting to higher-order patterns in the environment. Similar evidence comes from a color search task reported in Thomson et al (2012): when the probability of a repeated color was high given recent experience, the reaction-time cost of processing a switch trial (in which the color was not repeated from the previous trial) was high; when repetition probability was low, switch cost was also lower. Jones and Sieck (2003) find that sensitivity to sequential dependencies also affects more complex behaviors such as categorization, where participants accurately learn repetition probabilities of outcomes.…”

A Resource-Control Account of Sustained Attention

“…As the monitor learns that only a small proportion of trials require that he or she implement a course of action (whereas the vast majority of trials do not require the monitor to do anything), he or she may adjust the level of executive control (or attentional focus) to be more consistent with the most commonly occurring trial type. Indeed, probability-based modulations in performance have been shown to exist in standard performance tasks, such as Stroop (Logan & Zbrodoff, 1979) and even simple search (Thomson, D’Ascenzo, & Milliken, 2013), and a role for learning has been suggested for the deployment of efficient cognitive control (Verguts & Notebaert, 2009). In other words, executive control processes may be reduced over time simply because the observer learns to engage in less effortful processing.…”

Implicit learning modulates attention capture: evidence from an item-specific proportion congruency manipulation

“…Additionally, when many prior experiences are contextually similar to the current one, priming effects are dependent on the number of intervening experiences that occurred between the current and “influencing” trial, which has been interpreted by some as evidence for a form of retrieval interference (Thomson and Milliken, 2013b; Experiment 3). It has also been shown that the benefits and costs of target feature repetitions and alternations vary depending on whether one is in a search context in which target repetitions are likely or unlikely (Geyer and Müller, 2009; Thomson et al, 2013). …”