Mutual Interference Between Statistical Summary Perception and Statistical Learning

Abstract

The visual system is an efficient statistician, extracting statistical summaries over sets of objects (statistical summary perception) and statistical regularities among individual objects (statistical learning). Although these two kinds of statistical processing have been studied extensively in isolation, their relationship is not yet understood. We first examined how statistical summary perception influences statistical learning by manipulating the task that participants performed over sets of objects containing statistical regularities (Experiment 1). Participants who performed a summary task showed no statistical learning of the regularities, whereas those who performed control tasks showed robust learning. We then examined how statistical learning influences statistical summary perception by manipulating whether the sets being summarized contained regularities (Experiment 2) and whether such regularities had already been learned (Experiment 3). The accuracy of summary judgments improved when regularities were removed and when learning had occurred in advance. In sum, calculating summary statistics impeded statistical learning, and extracting statistical regularities impeded statistical summary perception. This mutual interference suggests that statistical summary perception and statistical learning are fundamentally related.

Towards a theory of individual differences in statistical learning

“…Indeed, there is compelling evidence that learners can exploit more than one source of statistical information at the same time (e.g. [55,56,59]), although sometimes at the cost of interference [60]. 5 Because sensory information related to SL phenomena is mostly visual or auditory, the tactile modality is omitted for the sake of simplicity.…”

What's statistical about learning? Insights from modelling statistical learning as a set of memory processes

“…Behaviourally, these two types of statistical learning appear to operate under different constraints (e.g. [26,28,29]), and to mutually interfere with each other such that detecting one kind of structure impairs detection of the other kind [30]. Perhaps related to or emerging from this interference, the time course of the detection of conditional and distributional regularities also differs.…”

Task specificity of attention training: the case of probability cuing

“…On the other hand, studies have shown that the training task sometimes modulates what is learned. For example, when performing a task that extracts the “summary statistics” of an array of lines (e.g., the average orientation), participants fail to learn the co-occurrence of lines on that array (Zhao, Ngo, McKendrick, & Turk-Browne, 2011). In addition, spatial context learning of a visual display shows limited or no transfer between visual search and change detection tasks (Jiang & Song, 2005).…”