11.10.2023; Vortrag
Q. Huang: Spontaneous information organization in Working memory
Max Planck Institute for Human Cognitive and Brain Sciences
Department of Psychology
Abstract
Working memory (WM) is a core cognitive function to flexibly inform and guide future behavior, with its capacity constrains various cognitive abilities. In real life, individual item in a complicated scene is usually not independently from each other, but depends largely on the background knowledge and structure our visual system brings to bear on it. For example, when remembering a telephone number, we not only need to memorize all the numbers, but also the ordinal relationship between numbers. The ability to encode and represent the temporal order of discrete elements occurring in a sequence is defined as sequence memory. To answer how a sequence is represented and stored in WM, in our tasks, we asked participant to memorize a sequence of item in combination with EEG/MEG recording. We find that serially remembered items are successively reactivated in a reverse order, and this backward replay pattern is highly related to behavioral recency effect. These findings consistent with observed neural replay in episodic memory, decision making and so on, reveals general neural mechanism for mental organization of external information. Meanwhile, daily experiences also involve the processing of multiple sequences, yet storing them may challenge the limited capacity of working memory. Therefore, in our recent study, we aim to investigate whether relational structures shared by multiple sequences would be spontaneously leveraged to reorganize and compress multiple information and its neural mechanism. We demonstrate that shared structure facilitates WM of multiple sequences and improve memory performance. Interestingly, when asking participant to recall one sequence, the other sequence sharing the same structure undergo spontaneous forward neural replay. These results suggest that replay-based neural mechanisms in WM mediate sequence combinations based on common structures.