A guess today may be a strategic error tomorrow: Predicting intra-individual differences in visual working memory

Abstract

Inter-individual differences in visual working memory (VWM) are well documented, however, there has been less work exploring intra-individual differences or how VWM processes may fluctuate over time in healthy young adults. It has also been documented that VWM is impacted by changes in female sex hormones in menopausal woman, however, these hormones also fluctuate across the monthly cycle in young woman, granted to a lesser extent. Given the high prevalence of female participants in the majority of VWM studies, understanding whether there are in fact intra-individual changes in VWM performance over the monthly cycle may not only provide a novel contribution to the discussion of fixed vs. flexible capacities in VWM, but may also help explain unreplicable results or inconsistencies in the literature. Here, we compared performance on a VWM colour-wheel task in women at two time-points (at menstruation and ovulation). Employing the 3-component swap model of VWM (Bays, Catalao, & Husain, 2009), we estimated measures of probability of target, non-target (swaps between target and non-target items), and guess responses, addressing the question of whether VWM performance varies over time within individuals and whether this is due to changes in the number of target responses or the types of errors made. Interestingly the probability of target response did not differ between the two time-points, however, the types of errors did. When VWM is at capacity (load 4 and 6) women made more swap errors at ovulation and more guesses at menstruation. These results suggest that the amount of information held in VWM may increase at ovulation but this results in more interference, with no concomitant increase in target responses. These results reveal the presence of intra-individual differences in VWM performance (types of errors made) that follow cyclical patterns in the monthly hormone cycle in young woman. Meeting abstract presented at VSS 2015