Abstract
When a decision task (T2) follows a working memory (WM) sample presentation, some studies reveal retroactive interference with memory consolidation. This effect contradicts earlier research that reported proactive interference in similar tasks. It is unclear what aspects of T2 lead to retroactive interference instead of or in addition to proactive interference, motivating the current goal of establishing boundary conditions for retroactive interference—i.e., under what circumstances WM consolidation can be interrupted. The present research investigates these boundary conditions via a series of modified versions of the WM/decision dual-task paradigm. While pilot versions of some experiments were previously presented, all current experiments include expanded samples and reanalysis using a signal-detection measure of memory strength, which removes contributions of decision bias changes that can contaminate an accuracy- or k-based analysis (c.f. Williams et al., 2022, https://doi.org/10.1037/xhp0001055). In one experiment, when T2 was a further WM sample, no retroactive interference was obtained. In another experiment, both a typical decision T2 condition and a “neutral” T2 that required interpretation without a motor response were included; only the typical T2 led to clear retroactive interference. There was weak evidence suggesting equivalent performance between the neutral and single task conditions. However, there was insufficient evidence that neutral performance differed from the dual task condition, meaning that the status of the neutral T2 may require further investigation. In a third experiment, T2 required only detection, placing minimal demands on decision processes. Surprisingly, mere detection still yielded retroactive interference that was comparable to that in a separate group of participants using a typical T2. Integrating across experiments, it appears that a task switch (from WM encoding to another task) with an immediate speeded motor response is sufficient to interrupt WM consolidation.