Abstract: PURPOSE: The interaction between working memory (WM) and perception has been explored extensively in cognitive psychology and neuroscience. However, the mechanisms of WM-perception interactions are not clear. We argue that the recent finding of a laminar-specific circuitry in the visual cortex of the human brain, enabling efficient cortical-layer interplay during information processing, casts unique light on the exploration of the WM-perception interaction and inspires novel predictions. This study examined the cortical-layer interplay account of WM-perception interaction by taking the phenomenon that visual WM load impairs visual detection (short for WM-iD below) as an example.
METHODS: Experiment 1 took advantage of the rational resource allocation of WM, testing the influence of set size of memory array on WM-iD. We required participants to memorize one, four, or eight colors. The cortical-layer interplay account predicts that the visual detection performance stopped decreasing when the allocated WM resource was at plateau, yet no detection-dropping occurred if the task only contained feed forward signals. That is, the visual detection performance is impaired in the 4-color condition compared to the 1-color condition but remains stable between the 4-color and 8-color conditions. Contrastingly, the sensory load account predicts that visual detection performance gradually decreases as memory load increases. Experiment 2 tested the influence of feedback signal of WM storge in WM-iD. We used the same parameters as in Experiment 1, but required participants to judge how many distinct colors appear, such that only the feedforward signal existed. Experiments 3 examined a novel prediction of the cortical-layer interplay account: visual WM and visually presented verbal WM load similarly affect visual detection, which is against the prediction of load theory. We required participants to memorize the colors (visual memory) or the letters (verbal memory) of one, six, or eight colored letters.
RESULTS: In line with the cortical-layer interplay account, the visual detection performance stopped decreasing when the allocated WM resource was at plateau (Experiments 1), yet no detection-dropping occurred if the task only contained feedforward signals (Experiments 2). Besides, visual WM and visually presented verbal WM load similarly affect visual detection (Experiments 3).
CONCLUSIONS: These findings support the cortical-layer interplay account to the underlying mechanism of the visual WM load impairing visual detection. Therefore, laminar-specific circuitry should be considered a factor in exploring WM-perception interaction in the future.

Key words: working memory, perception, visual detection, laminar-specific circuitry