High value leads to improved explicit recollection, but reduced implicit memory, when learning kaleidoscope images
Michael S. Cohen1, Larry Cheng1, Ken A. Paller1, Paul J. Reber1; 1Northwestern University
Recognition judgments typically rely on MTL-dependent explicit memory, but in some cases, participants can make accurate guesses seemingly driven by implicit memory (e.g., Voss et al. 2008). These “implicit recognition” effects provide a rare opportunity to observe interactions between implicit and explicit memory systems, but have also proven somewhat challenging to study. Here, we use a value-directed recognition (VDR) paradigm to elicit recognition responses that either depend on explicit memory or benefit from a contribution of implicit memory. Two novel kaleidoscope images were shown simultaneously in two different spatial quadrants during study, one marked high-value and one low-value, in blocks of 16 stimuli. After each study block, participants performed source memory (quadrant) and yes-no recognition to assess explicit memory. Explicit memory was better for high point-value items, likely due to increased encoding effort. After all 7 study blocks, participants also completed a forced-choice recognition memory test, previously shown to be more sensitive to contributions of implicit memory. Here, there was an interaction between value and confidence, with low point-value items tending to produce higher levels of accurate guessing on the forced-choice test than high-value items. Thus, previously reported effects of VDR—improved explicit memory for high-value items—extend from verbal memory to novel, non-verbalizeable stimuli, while more robust implicit memory contributions can be elicited from items marked low-value. This approach will enable further investigation of the neural basis of interactions between these two forms of memory, and of when implicit memory representations inform guesses on recognition memory tests.