The Academic Salon of the Department of Psychology and Cognitive Sciences is held every Thursday afternoon. Welcome to all students and faculty members from every department!
Time: May 22 (Thursday) afternoon 14:30
Location: Room 1110, 11th floor, Lv Dalong Building
Department of Psychology and Cognitive Sciences MI Yuanyuan's Research Group Academic Salon
Report One
A Dual-Process Model of Serial Dependence Effects
Presenter: Xiuning Zhang
Content:
Past experiences automatically shape perception, a phenomenon that widely exists in various cognitive functions, reflecting the brain's utilization of temporal patterns in the environment. Among these, serial dependence is a typical example, reflecting the regulatory effect of recent sensory history on current perception, and manifesting as two opposite bias patterns: perception may be repelled or attracted by recent stimuli. Existing research shows that repulsive bias typically occurs during early sensory processing stages, while attractive bias appears more in post-perceptual decision stages. Current mainstream serial dependence models adopt a dual-process framework of efficient coding and Bayesian decoding to explain these perceptual biases. This model suggests that adaptation in early sensory processing stages makes perception move away from previous stimuli, while decision traces attract perceptual reports toward previous stimuli. This framework reveals that different mechanisms operate separately at different stages of information processing, jointly promoting the separation (repulsion) and integration (attraction) of temporal information. What are the underlying neural and computational mechanisms?
Report Two
Storage in Visual Working Memory Recruits a Content-Independent Pointer System
Presenter: Changhan Lin
Content:
Previous research has shown that information storage in working memory triggers stimulus-specific neural activity that can track stored content. This study reveals a unique class of load-sensitive neural activity that can mark items but does not itself represent the specific content of these items. Researchers recorded electroencephalography (EEG) activity when adult human participants stored different numbers of items in visual working memory. Through multivariate analysis of scalp potential topographic maps, they achieved precise tracking of the number of items stored by individuals and effectively predicted individual differences in working memory capacity. Crucially, this neural signature of working memory load remained stable across different visual feature types and numbers of stored visual feature conditions, indicating that it tracks the quantity of individual memory representations rather than memory content. The researchers hypothesize that these findings reflect a capacity-limited pointer system that supports online storage and attentional tracking.
Report Three
Attention Modulates Segmenting and Chunking in Continuous Temporal Sequence Processing
Presenter: Zhitao Feng
Content:
Dividing temporal sequences into discrete events is a fundamental mechanism of brain information processing. Our nervous system dynamically parses continuous sensory input into a series of discrete events (such as chunks) through attention. These events have hierarchical structures and form scaffold-like neural representations. This event segmentation mechanism not only significantly reduces cognitive load but also enhances the brain's processing flexibility in multitasking and dynamic environments. So, what are the neural and computational principles behind this mechanism?
