Comparison Processes are Necessary and Sufficient for Memory Bias Persistence

Humans constantly compare their visual memories with perceptual inputs to modify their behavior adaptively. However, basic and applied studies suggest that these comparisons can produce systematic biases in observers’ memory reports that persist across time and contaminate subsequent reports. This has led to the hypothesis that perceptual comparisons, while adaptive, are both necessary and sufficient to produce persistent memory biases. In two experiments, we directly tested this hypothesis by independently manipulating demands for perceptual comparisons and memory reports across a pair of memory tests separated by a 24-hour delay. At each test, observers were cued to recall colored object silhouettes that they had learned previously. At the first test, following cued recall, the silhouette was presented again in a unique color to serve as a perceptual probe to be compared with memory or to be treated as a task-irrelevant distractor. Afterwards, the observer either reported their memory of the recalled silhouette or not. Observers’ memory reports at the first test contained attractive biases towards the probe immediately after comparisons and distraction. However, the persistence of these biases was selective. Biases induced by comparisons, but not distraction, persisted from the first to second test, suggesting that comparison processes were necessary for persistence to occur. Moreover, this persistence occurred even when observers did not report their memory immediately after the comparison, suggesting that the comparison alone was sufficient. Together, these findings suggest that processes tied to perceptual comparisons are mechanistically responsible for the persistence of memory biases over time.