Information Transmission In VHPC-mPFC Sparse Causal Network During Working Memory

ObjectiveWorking memory(WM)is one of the important cognitive functions.Hippocampus(HPC)and prefrontal cortex(PFC)are responsible for working memory.The information transmission between HPC and PFC is one of frontier science issues in the study of working memory mechanism.Taking the working memory of rats in the Y-maze as research objects,the present thesis developed a sparse causal network analysis method based on the neural ensemble rate coding and built the HPC-PFC sparse causal network of action potentials(AP).The bidirectional information flow between HPC and PFC was calculated.It is expected to provide new insights and methods for the information transmission mechanism of working memory in brain networks.Methods1.Animals: Sprague-Dawley(SD)rats(male,12-14 weeks,300-350g).2.Working memory behavior in Y-maze: delayed-alternation task,the behavioral indicators(the correct percent and latency)were recorded during the working memory task.3.Experimental data: in vivo implantable electrodes recording technique was employed for awake animal to record the neuronal activity(16-channel)from rat HPC and PFC when they:(1)correctly performed the working memory task,(2)incorrectly performed the working memory task,(3)at resting state.Neural signals were high-pass filtered(>500Hz).Pre-set voltage thresholds was used for spike detection.Spatio-temporal trains of action potentials from HPC and PFC neurons were then obtained after spike sorting.4.Sparse space of action potential was built based on the neural ensemble rate coding.5.Construction of sparse causal network based on action potentials: The maximum likelihood estimation causal analysis method was used to estimate the causal connection among action potentials to build three sparse causal network(HPC,PFC and HPC-PFC).The sparse causal network connection strength was calculated,respectively.6.The directional information flow between HPC and PFC was calculated,which quantitatively describe the magnitude of information transmission from HPC to PFC and from PFC to HPC.7.Statistical analysis:(1)One-way analysis of variance was used to assess: the behavioral indicators during the working memory task training;the changes of HPC causal network connection strength over time;the changes of PFC causal network connection strength over time;The information flow between HPC and PFC over time.(2)Paired sample t-test was used to compare: causal network connection strength in HPC and PFC during working memory;the bidirectional information flow between HPC and PFC.(3)Independent sample t-test was used to compare: the information flow between HPC and PFC in correct trials and at resting state;the information flow between HPC and PFC in correct trials and incorrect trials.Results1.Behavioral results:(1)Correct percent: during the working memory task training,the averaged correct percent gradually increased from 68.38±6.24%(day 1),to 92.22±2.45%(day 5),92.86±3.05%(day 6),which met the working memory training standard.(2)Latency: during the working memory task training,the averaged latency gradually increased from 17.42±3.97s(day 1),to 3.07±0.23s(day 6).2.Construction of sparse space based on action potentials: action potential trains from 6 rats(16,21,18,19,19 and 17 neurons from HPC,15,17,21,21,15 and 18 neurons from PFC)were obtained.Action potential trains from 5 neurons were raised to construct the sparse space.3.The sparse causal network connection strength of HPC and PFC during working memory:(1)The sparse causal network connection strength of HPC and PFC peaked prior to the work memory behavior reference point(RP).(2)The sparse causal network connection strength in HPC was significantly greater than that in PFC,and the peak of connection strength in HPC was prior to that that in PFC(P < 0.05).4.The directional information flow in HPC-PFC sparse causal network during working memory:(1)The directional information flow peaked prior to the RP.(2)The information flow from HPC to PFC was significantly greater than that from PFC to HPC,and the information flow from HPC to PFC peaked prior to that from PFC to HPC(P < 0.05).5.The directional information flow in the correct trials was significantly greater than that in the incorrect trials and at resting state.Conclusions1.Brain network mechanism for working memory:There are neural information(action potetial)networks in HPC and PFC for working memory.The peak of network connection strength is prior to the behavior reference point.The network connection strength in HPC significantly greater than that in PFC.These results indicated that the hippocampus network is the main network for working memory in rats(rodents).2.Information transmission mechanism for working memory:There are bidirectional information transmissions between HPC and PFC characterized by the information flow for working memory.The peak of information flow is prior to the behavior reference point.The information flow from HPC to PFC peaked prior to that of reverse transmission.These results indicated that the information transmission from HPC to PFC might be the main information transmission direction for working memory in rats(rodents).3.The necessity of information transmission between HPC and PFC for working memory:The information flow on the correct trials was significantly higher than that on the incorrect trials and control conditions.These results indicated that the information transmission between HPC and PFC is necessary for working memory.