A Study On Characteristics Of Functional Network Constituted With Action Potentials In Neural Ensemble During Working Memory Task

Objective:Working Memory (WM) is an important cognitive function, research on functional connectivity of brain network during WM is one of a major breakthrough of understanding the mechanism of WM. In order to highlight the characterization of functional connectivity among neural information in WM, neural assemblies were selected from the high-dimensional original signals (Action potentials, Aps) based on the rate coding. Then, a low-dimensional neural ensemble (NE) space was constructed by neural assemblies. Under the low-dimensional space, the value of the causal connectivity and the efficiency of information transmission in causal network were calculated to provide support for describing the brain network in WM quantitatively and effectively.Methods:1. Experimental data:In the process of adult SD (Sprague-Dawley) rats working memory in Y-maze, multi-channel implantable electrodes array were applied to record multi-channel neural electrical activities in the medial prefrontal cortex, through high-pass filter, peak potential (spike) detection and classification processing (spike-sorting), to obtain the neurons’action potential (APs). The data come from the WM of4rats,10trails per rat.2. Rate coding of neural assembly:Selecting a moving window (window width is500ms, moving step is125ms), counting the spiking rate of neurons within each window, calculating the distribution of rate coding.3. Selecte the neural ensemble and constructe low-dimensional NE space: Selecte the neuron which has higher spiking rate than average spiking rate of all neurons within2s before RP in each trial as a member of neural ensemble. Then, a low-dimensional NE space was constructed by neural ensemble.4. Causal connectivity among APs based on maximum likelihood estimation (MLE):Using the method of MLE, the causal connectivity matrix and the average of causal matrix (Cc) were calculated in the low-dimensional NE space, in order to quantitatively describe the functional connectivity among APs. 5. The efficiency of information transmission in brain network of WM:Definite a causal network based on the value of causal connectivity. The global efficiency of network was calculated to quantitatively describe the efficiency of information transmission in causal network.Results:1. Neuron’s action potentials during rat WM task in Y-maze.Action potentials of neurons were sorted out from each rat:19neurons for rat1,23neurons for rat2,25neurons for rat3,20neurons or rat4.2. Rate coding of of neural ensemble during WM.There were obvious neural ensembles within2seconds before reference point (RP) for4rats during WM tasks.3. Construction low-dimensional NE space.Ratl had4neurons in neural ensemble, the proportion of the neurons in the NE space to the number of all the neurons (19) is21.05%. Rat2had6neurons in neural ensemble, the proportion of the neurons in the NE space to the number of all the neurons (23) is26.09%. Rat3had5neurons in neural ensemble, the proportion of the neurons in the NE space to the number of all the neurons (25) is20%. Rat4had6neurons in neural ensemble, the proportion of the neurons in the NE space to the number of all the neurons (20) is30%. Respectively, the meurons in neura ensembles of each rat were used to construct low-dimensional NE space.4. Causal connectivity among APs during WM tasks.In the low-dimensional NE space, the value of Cc was changing dynamically and increased to maximum before RP during WM tasks for each rat (the value of Cc increased to maximun within2s bfore RP for ratl and rat2. the value of Cc increased to maximun within Is bfore RP for rat3and rat4). The value of Cc in ratl was increased from initial period (pre)(1.0745±0.3811) to peak (2.5465±0.3617)(t-test, p<0.05). The value of Cc in rat2was increased from pre (0.5974±0.1132) to peak (1.0047±0.0917)(t-test, p＜0.01). The value of Cc in rat3was increased from pre (0.6436±0.1752) to peak (1.4054±0.1829)(t-test, p＜0.01). The value of Cc in rat4was increased from pre (0.8527±0.2496) to peak (2.6502±0.5394)(t-test, p＜0.01). Compared with the value of Cc in the original space, the peak value of Cc was significantly larger in the low-dimensional NE space (t-test, p＜0.001).5. Global efficiency (Eglob) of causal network among APs during WM tasks.In the low-dimensional NE space, the value of Eglob was changing dynamically and increased to maximum with the same period of Cc. The value of Eglob in ratl was increased from pre (0.2334±0.0926) to peak (0.3994±0.0533)(t-test, p<0.05). The value of Eglob in rat2was increased from pre (0.2167±0.0441) to peak (0.4195±0.0272)(t-test, p＜0.001). The value of Eglob in rat3was increased from pre (0.1357±0.0384) to peak (0.3746±0.0488)(t-test, p<0.01). The value of Eg,ob in rat4was increased from pre (0.2098±0.0617) to peak (0.4861±0.0339)(t-test, p<0.01). Compared with the value of Eglob in the original space, the peak value of Eglob was significantly larger in the low-dimensional NE space (t-test, p＜0.001).Conclusion:1. Neural ensemble was existed in WM, the sparse spiking is less than or equal to30%.2. Functional connectivity among APs was enhanced (Cc↑) and the efficiency of information transmisson in APs network was enhanced (Eglob↑) during WM.3. The characteristics of brain network connectivity of neural ensemble can be used to eminently describe the WM.