I focus more on the first two projects now~
How does hippocampal ripples and hippocampal-cortical communications support visual short-term memory and long-term memory? (master project, 2022/03/24 - now)
Supervisor Associate Research Fellow Ying Cai, Research Assistant Professor Jing Liu
Proposal Hippocampus is a subcortical node underlying memory consolidation. Extant evidence suggested hippocampal ripples and hippocampal-cortical communication contributed to human declarative memory. However, it remains largely unknown 1) whether visual short-term memory and associative memory also engage such neural substrates, 2) the contribution of hippocampal-cortical communication to the dynamics of visual short-term memory and associative memory. Using stereo-EEG recordings, we have been examining whether the hippocampal ripples and hippocampal-cortical communications support the transformation of visual representations from encoding to short-term memory maintenance to long-term Memory retrieval.
Progress completed ripple rates analysis, working on ripple-induced communication and representation analysis (2023/03/03)
Distinct and Shared Neural Correlates of Quantity and Quality in Visual Working Memory (master project, 2021/09/30 - now; WMS talk, 2023/06/20)
Supervisor Associate Research Fellow Ying Cai, Professor Shaomin Zhang
Abstract Visual working memory (VWM) is fundamental for human cognition, but it has limited quantity and quality. In this study, we aim to investigate whether memory quantity and quality have distinct or shared neural correlates with intracranial-EEG signals from 11 epilepsy patients. By employing unsupervised clustering analysis, we found that in low-frequency neural activity (3-12 Hz), the brain represented memory quantity only during VWM maintenance, while in high-frequency (70-140 Hz), quantity-sensitive correlates existed during both encoding and maintenance. Additionally, by correlating both low- and high-frequency neural activities with memory error at the trial level, we identified quality-sensitive correlates and found that quantity and quality overlapped in both frequency bands. Intriguingly, focusing on a series of memory-related brain areas, we found that areas with higher memory representational dimension (from visual, parietal, frontal cortex, to medial temporal lobe) relied more on low-frequency neural activity to represent quantity while relying more on high-frequency neural activity to represent quality. And only frontal and parietal had channels representing quantity and quality simultaneously. Together, both data-driven and theory-driven analyses demonstrate that quantity and quality show distinct and shared neural correlates, and they may be dissociated VWM components receiving shared feedback from cognitive control network.
Progress collected 11 patients data, completed major analysis, presented the work at WMS, plan to do some control analysis (2023/07/07)
How does parietal-hippocampus communications support memory binding? (master project, 2021/09/30 - now)
Supervisor Associate Research Fellow Ying Cai, Professor Shaomin Zhang
Proposal Hippocampus plays an important role in binding memory information. Neocortex, such as parietal cortex, is considered representing memory information. However, it remains unclear how hippocampus interacts with parietal cortex to support memory binding and their dialogue may vary in different stages. For example, during the encoding stage, parietal cortex may encode memory information through conjunctive encoding, which is projected to hippocampus to support mnemonic binding. When provided memory cues, hippocampus retrieves the memory via pattern completion and communicates it to parietal cortex to reactivate corresponding memory representation. Using stereo-EEG recordings, we plan to directly examine the above hypothesis.
Progress collected 6 patients data, waiting for more patients, plan to build the analysis pipeline (2023/03/03)
Does Imagery share similar neural mechanisms with the actual movement? (NMA project, 2022/07/11 - 2022/07/29)
Does confidence and uncertainty differ in visual working memory? (undergraduate senior thesis project, 2021/01 - 2022/06; WMS talk, 2022/06/23)
Progress conducted a control experiment, writing
Spatial processing mediates the effect of electrical stimulation over posterior parietal cortex on visual short-term memory (tES project, 2020/12 - 2022/04; CNS poster, 2022/04/24)
Abstract The causality between posterior parietal cortex (PPC) and visual short-term memory (VSTM) is still controversial. Accumulating studies found PPC played a critical role in spatial processing, and we examined whether spatial processing mediates the function of PPC in VSTM. Using within-subject design (n=30), we explored whether and how anodal electrical stimulations over PPC affected recall performance in delay estimation tasks for locations and colors. Every three days, subjects accepted PPC stimulation (P4 electrode in 10-20 EEG recording system, 2mA, 20min), occipital stimulation (active control: Oz, 2mA, 20min), and sham stimulation (passive control: half subjects in P4 and half in Oz, 2mA, 30s). Before and after each stimulation, subjects completed delay estimation tasks for locations and colors (set size = 8), and the 3-factor mixture model was used to estimate the recall precision, the probability of recalling target(pT), non-target (pNT), and random guessing (pU), respectively. Our results revealed, in the location WM task, the PPC stimulation decreased the random guessing compared with occipital/sham stimulation (ps < 0.024), but didn’t change recall precision (ps > 0.444). In contrast, in the color WM task, the PPC stimulation increased the recall precision compared with the sham condition (p = 0.04; compared with the occipital stimulation: p = 0.20), but didn’t change random guessing (ps > 0.170). Moreover, the interaction effects between tasks and PPC-sham differences were significant (compared with sham: ps <0.03). Together, our results suggest PPC plays a general casual role in VSTM, but its specific function is mediated by spatial processing.