Inik Kim¹, Bo-Yeong Won², Ana Acebo¹, Weiwei Zhang¹; ¹University of California, Riverside, ²California State University, Chico

The ability to remember detailed information is critical not only for daily functioning but also for understanding memory declines with aging and neuropsychiatric disorders, such as schizotypy. Given the hippocampal pattern separation mechanism for Long-Term Memory (LTM) precision, the interactions between Short-Term Memory (STM) and LTM by Cowan’s embedded process model, and the growing literature on the hippocampus’ roles in STM, we hypothesized that hippocampal pattern separation as a neurocomputational mechanism for mnemonic precision in visual STM would limit the task performance when maintaining precise visual information concurrently in LTM and STM. Specifically, the high task demand on pattern separation in LTM would degrade mnemonic precision for a concurrent visual STM task. To test this prediction, we embedded a Change Detection (CD) task into the Mnemonic Similarity Task (MST) test phase. Participants were required to perform the MST test phase during the delay interval of the CD with line orientations. Experimental manipulation (high-capacity load versus high-precision load, Experiment 1) and ROC method (Experiment 2) were used to tease apart WM quantity and precision. The results from both experiments indicated that memory precision was selectively impaired in the MST lure trials that required the ability to remember the rich details of the item, whereas memory quantity remained intact. Experiment 1 specifically demonstrated that CD accuracy was notably impaired during the MST lure trials under the high-precision load while remaining intact for other MST trial types and under the high-capacity load of the CD task. Experiment 2 revealed that the ROC-based computational estimate of mnemonic precision, but not capacity, was impaired during the MST lure trials. These findings support the hippocampal pattern separation hypothesis by demonstrating that hippocampal pattern separation is pivotal for visual STM precision.