The amount of information that can be stored online in working memory (WM) is of central theoretical and practical importance in understanding the limits of our cognitive abilities. Despite the importance of WM in our daily lives, it has a surprisingly limited capacity 5, 6, 7, 8.

This capacity is involved in nearly all domains of cognition and its central importance is reflected by its high correlation with measures such as fluid intelligence 1, 2 and academic success 3, 4. Working memory (WM) refers to the capacity to keep information active and accessible when it is no longer present in the environment. Crucially, standard discrete response paradigms underestimate the amount of information in memory representations.

This paradigm provides multiple pieces of evidence that memory representations are rich and probabilistic. The later bets contained target information even when the first response would generally be classified as a guess or report of an incorrect item, suggesting that such failures are not all-or-none. This resulted in memory performance improvements when averaging across bets, and overall suggests that memory representations contain more information than can be conveyed by a single response. While the first bet was on average the most precisely placed, the later bets systematically shifted the distribution closer to the target, resulting in asymmetrical distributions about the first bet. Bet dispersion correlated with performance, indicating that internal uncertainty guided bet placement. Rather than collecting a single report per trial, we had participants place multiple bets to create trialwise uncertainty distributions. When asked to remember a color, do people remember a point estimate (e.g., a particular shade of red), a point estimate plus an uncertainty estimate, or are memory representations rich probabilistic distributions over feature space? We asked participants to report the color of a circle held in working memory.

Her betperformance is rated