Cognitive Brain Network Based On SEEG

The brain’s cognitive function is the basis for human understanding and transformation of the world,and the core of brain cognition is memory,whose generation mechanism and processing mode have received extensive attention from researchers.Electroencephalogram(EEG)signals have become a key method to the research of cognitive psychology,which is contributed by the millisecond time resolution and highly integrated spatial features.Yet most research has used cortical EEG with high noise content,which are difficulty in feature extraction and cognitive analysis.In addition,researchers often focus on a single functional area of the brain and lack in-depth studies at the whole brain level.At the same time,Stereoelectroencephalography(SEEG)is widely used in the fields of brain-computer interface and computational neuroscience because of its high signal-to-noise ratio and close to the real state of brain activity.Therefore,in this thesis,SEEG signals were collected in a cognitive memory experiment and a causal brain network of cognitive memory with synergistic effects of multiple brain regions was constructed and a study of the brain’s cognitive memory processes based on the dual dimension of time and space.The details of the study are as follows:(1)This thesis focuses on the spatial cognitive memory process,and construct a causal brain network oriented to mental rotation by adopting an experimental paradigm of mental rotation reflecting visual-spatial memory cognition.Firstly,SEEG signals of the mental rotation experiment and the behavioral performance were recorded,at the same time,Laplacian filtering and other pre-processing methods were applied to highlight the signal characteristics.Secondly,the functional connection brain network was constructed based on the time-frequency analysis of the behavioral performance and SEEG signals,which help to the investigation of the functional connections and active frequency bands of deep brain regions.Finally,granger causality analysis was applied to refine the connection direction of brain regions in the functionally connected brain network,thus innovatively constructing a causal brain network model oriented to mental rotation.The experimental results show that the frequency range of the cognitive process of mental rotation is mainly concentrated in the theta frequency band,and the nodes in the causal brain network model are mostly the right brain area,which indicated that the brain mainly relied on the right brain to realize spatial visual imagination during mental rotation,further confirming the hemispheric effect of cognitive processes based on mental rotation.The outflow nodes of the information in the brain network model were found to be cortical brain areas such as the right anterior lobe and the right frontal lobe,which flowed to the hippocampus-temporal lobe(memory function brain area)via the post-central brain area,suggesting that spatial cognitive memory processes activate motor-related brain areas in the brain.(2)This thesis focuses on working memory,the core of short-term memory,and uses a change-detection experimental paradigm consistent with the three-stage process of working memory to construct a causal brain network for the information retention period during working memory.Firstly,the SEEG signals of working memory and behavioral outcome recordings were acquired and recorded,while the SEEG signals were subjected to preprocessing operations.Secondly,a functional connectivity brain network with a frequency range of 4-30 HZ was constructed by combining the results of time-frequency analysis of the working memory information retention period.Finally,the granger causality analysis was used to refine the connection direction of the functional brain network,and the causal brain network model for working memory was innovatively constructed.The experimental results show that working memory has a load effect,and memory load is positively correlated with response duration and negatively correlated with response accuracy.The causal brain network model shows that the outflow nodes of information during working memory have diverse characteristics,including the amygdala,temporal lobe,anterior lobe and frontal lobe and other regions,and the information finally converges to the hippocampus-temporal lobe,which indicates the wide range of brain activation areas for information memory process.The causal brain network model reveals that the orbitofrontal lobe has a learning function,and information is transmitted through brain regions such as the frontal lobe,and information is remembered and stored in the hippocampus-temporal lobe region.