When repeated presentation of visual feature bindings does and does not result in learning: Visual short-term and long-term memory are distinct but work in tandem

Two experiments explored a previous finding that 120 repetitions of the same six-item array for change detection resulted in no, or very slow learning. This contrasts with learning from six repetitions tested by recall. Shimi and Logie (2019) proposed that array repetition for change detection is supported by a limited capacity, temporary visual cache memory, with contents replaced by the next study array, even if it is identical. This is coupled with a weak episodic trace that strengthens across repeated presentations but requires more than 60 repetitions for learning. Experiment 1 tested (1) whether a short (500 ms) study-test interval would result in reliance on the visual cache with no evidence of learning across 120 repetitions, and (2) whether a longer (5000 ms) study-test interval would gradually strengthen the episodic trace but require many repetitions for learning. A 500 ms study-test interval showed no learning after 120 repetitions, and participants reported being unaware of the repetition. A 5000 ms study-test interval showed performance improvements but only after 40 repetitions and participants reported becoming aware of the repetition. In Experiment 2 different arrays on each of 120 trials with short and long study-test intervals showed the same lack of learning found for the 500 ms study-test interval in Experiment 1. Results appear consistent with a limited capacity visual cache memory for change detection that retains the array only for the current trial, working in tandem with a weak episodic trace that accumulates across trials but only supports performance after multiple repetitions and longer study-test intervals.