Abstract
Our experiences contain both statistical regularities (e.g., the route and scenery on your walk home) and idiosyncratic details (e.g., the time you ran into an old friend on the sidewalk). Despite the commonalities among our experiences, we are able to form rich memories of these unique events. Prior work has demonstrated that attention is captured by and oriented to regularities (e.g., Zhao et al., 2013), yet attention also powerfully influences what we encode and subsequently remember. Together, these facts pose a paradox: How can we encode unique details of our environment if attention is drawn to the regularities in our environment? To address this question, we adapted temporal and spatial visual statistical learning paradigms to include: (a) trial-unique information embedded in the context of regularities and (b) surprise memory tests of the trial-unique information. In the temporal domain, we found that memory is suppressed for predictive items (e.g., first items in a temporal pair), suggesting that attentional orienting to predicted information may suppress the encoding of current information. In ongoing work in the spatial domain, we replicated previous work showing that attention is facilitated by regularities, as measured here by faster response times in structured displays. Additionally, we obtained preliminary evidence that memory is enhanced for information that co-occurs with regularities, suggesting that the attentional enhancement for regularities can spread over space. These data highlight interactions between statistical learning and episodic memory and suggest that these interactions may be mediated by attention.
Acknowledgement: NSF GRFP NIH R01 MH069456