Beijing, May 5 (Zhongxin Net) -- How humans remember and recall things is a difficult problem in brain science, in which working memory is a system for humans to temporarily save and use information, just like the high-speed memory of computers. It is not only necessary for daily life, but also the foundation of high-level cognitive functions such as language understanding, learning, and reasoning. So, how does the human brain process working memory? What are the features of this process?
In order to solve these problems, Jiang Tianzai's team from the Brain Network Group Research Center of the Institute of Automation of the Chinese Academy of Sciences (Institute of Automation, Chinese Academy of Sciences) cooperated with Zurich University Hospital in Switzerland to systematically analyze the functional division and collaborative mode of the human brain amygdala-hippocampal circuit in the working memory encoding and maintenance stage using intracranial EEG signals, a variety of EEG analysis methods and machine learning models.
The main findings of this published paper are the results. Photo courtesy of Institute of Automation, Chinese Academy of Sciences
They found that the human brain amygdala-hippocampal circuit showed obvious functional characteristics of "each performing its own duties" and "watching each other" in the working memory encoding and maintenance stage, and jointly revealed the neural representations and interaction patterns specific to the function-specific human brain amygdala-hippocampal circuit to support the processing of working memory encoding and maintenance stages. This not only provides a new perspective for the role of this circuit in working memory, but also has important significance for exploring the neural mechanism of working memory and understanding the principle of high-level cognitive function of the human brain.
This important achievement of brain science research paper was recently published online in the international academic journal Nature Communications. Co-first authors of the paper, associate researcher Li Jin and assistant researcher Cao Dan, Institute of Automation, Chinese Academy of Sciences, said that a total of 14 patients with refractory epilepsy collected intracranial EEG signals from the amygdala and hippocampus when completing working memory tasks, and used a variety of EEG analysis methods and machine learning models to analyze the neural representation patterns and information transmission directions of the amygdala and hippocampus in the working memory encoding and maintenance stage from the level of local activity and information interaction, and decoded the working memory load based on this.
It was found that the human brain amygdala-hippocampal circuit showed two obvious functional characteristics in the working memory encoding and maintenance stage: first, in terms of "each performing its own duties", the amygdala characterized the memory content in the working memory coding stage, while the hippocampus maintained the representation of memory information in the maintenance stage after the material disappeared. Characterization of the amygdala in the encoding phase and the hippocampus in the maintenance phase can better decode the working memory load.
Second, in terms of "watching each other", the amygdala and the hippocampus cooperate with each other in the encoding and maintenance phases to transmit information to each other, and the information flow emitted by the amygdala in the encoding phase and the information flow emitted by the hippocampus in the maintenance phase can better decode the working memory load.
Researcher Jiang Tianzai, co-corresponding author of the paper, said that this study revealed that the functional division of labor and collaboration mode of the amygdala and hippocampus support working memory processing by systematically analyzing the neural representation and information transmission patterns of the amygdala-hippocampal circuit in the working memory encoding and maintenance stage of the human brain. These findings not only provide a new basis for understanding the neural mechanisms of working memory, but also provide new perspectives for the study of related diseases, such as schizophrenia and Alzheimer's disease, which exhibit working memory deficits. In addition, by combining intracranial EEG recording, multivariate analysis, and machine learning methods, it also provides a new research framework for future brain-computer interface research and neurofeedback therapy. (End)