Abstract
Understanding how we consolidate visual information into working memory is critical for our understanding of any subsequent mechanism, such as memory storage or retrieval. However, here we suggest that the current, prominent understanding of how items are consolidated is incorrect. In particular, we find evidence against the influential claim that increasing the number of to-be-consolidated items slows the consolidation process. The time course of consolidation was examined in a novel paradigm in which an array of colored stimuli was presented for a variable amount of time. At the end of this stimulus period one item was immediately replaced with a mask, which also served as a cue, drawing attention to that item in the same moment it was masked. Participants then reported the color of the cued/masked item, and using a mixture model analysis we found that only a few items were available in memory after the cue/mask. This limited capacity existed even though there was effectively no memory retention interval, and even when the stimuli were presented for an entire second. Critically, however, shortening the stimulus period from 1000 ms to as little as 100 ms had almost no effect on performance. Moreover, although further shortening the stimulus duration below 100 ms did hurt performance, these effects did not interact with set size, implicating visual masking rather than an interruption of the consolidation process. We suggest that the previous demonstration of such an interaction did not in fact reflect encoding processes. Instead, we show that this result reflects a surprising effect whereby memory for masked items in a multi-item display continues to decay after the mask. These results suggest that visual encoding is capacity limited, but extremely fast and operates in parallel across items.
Acknowledgement: NIH NIMH Grant R15MH113075