工作记忆中持续性发放的机制研究

Working memory is a crucial brain process that temporally stores and manipulates information for active control of cognitive behavior. Delay related persistent activity has been observed in prefrontal cortex (PFC) in working memory and was thought to mediate the maintenance and manipulation of information. Currently two models are competing to explain the mechanism for persistent activity in working memory. One is the intrinsic excitability hypothesis, which claims that positive feedback mechanism of ion channels can lead to self-sustained plateau potential and persistent firing in the absence of sensory inputs. The other is recurrent network theory, which stresses the importance of synaptic recurrent connections among local and global networks in maintaining the positive feedback for persistent activity. Intrinsic excitability hypothesis predicts that temporarily enhancing or suppressing neural excitability by depolarizing or hyperpolarizing current injection will further enhance or suppress the excitability, therefore boost or decrease the subsequent firing even after ceasing of current injection, respectively. In contrast, recurrent network hypothesis predicts that perturbing single neurons' excitability will not modulate the subsequent firing. The two hypotheses also differ in prediction of synaptic noise in delay period, effects of perturbing global excitability, and sensitivity to pharmacological manipulations. We will test these predictions by combined efforts from in vivo whole-cell recording, extracellular recording, optogenetic, and pharmacological means in head-fixed mice performing working memory task. We have already measured membrane potential by whole cell recording while mice were performing a delayed non-match to sample task. The membrane potential modulation during delay period was measured and compared to neuronal firing. The results show that neuronal firing was sensitive to small fluctuations in membrane potential, as predicted by recurrent network hypothesis. We applied depolarizing or hyperpolarizing currents during delay period, but failed to observe autonomous modulation in membrane potential and firing rates following the ceasing of current injection. In future, we will further collect more neurons to corroborate our results. We will also use optogenetics to manipulate global neural excitability to test the two hypotheses. Furthermore, we will apply pharmacological method to block calcium activated non-selective cation channels (CAN) and L-type calcium channels, which are predicted to be critical by intrinsic excitability hypothesis but not by recurrent network hypothesis. Our results will illustrate which of the two hypotheses is closer to the reality of the underlying mechanism for persistent activity in working memory.

前额叶皮层在工作记忆任务的延迟期间存在持续性发放，被认为是大脑存储和操纵信息的媒介，但其产生机制不明。目前有"单细胞兴奋性"和"网络回环连接"两种假说可以解释持续性发放，并且具有明确而不同的预期实验结果。我们在小鼠中开发了基于嗅觉的工作记忆行为范式，将综合应用在体全细胞记录、在体胞外记录、光遗传、药理学等手段，尝试验证哪种假说更接近于生物学现实。我们将观察单个前额叶皮层神经元在工作记忆的延迟期间的膜电位，从而了解持续性发放中的突触输入特性；通过直接注入去极化或超极化电流，我们将观察撤销电流后神经元持续性发放的可能调控，从而直接验证两种假说预期正确与否；通过光遗传实验，我们将操控群体神经元的兴奋性，并观察对后续发放的影响；通过药理学研究，我们将特异性的阻断单细胞假说中的核心离子通道，从而验证假说的合理性。本项目将有助于阐明工作记忆中持续性发放产生的机制。

前额叶皮层在工作记忆任务的延迟期间存在持续性发放，被认为是大脑存储和操纵信息的媒介，但其产生机制不明。针对可以解释持续性发放的两种假说，本项目开展了一系列研究工作，实验结果支持其中的“网络回环连接”假说，而否定“单细胞兴奋性”假说。本项目加深了对工作记忆中持续性发放产生机制的理解。.主要研究内容和结果：（1）建立了小鼠嗅觉认知任务自动高通量训练系统 ，可以高效训练小鼠学习完成基于嗅觉的认知任务。在该系统中完成了小鼠延迟非配对等超过5种认知任务的训练，并与神经活动操纵和电生理记录、核磁共振成像技术整合；在嗅觉任务的基础上，还扩展了空间工作记忆的范式。（2）发现前梨状皮层参与工作记忆。选择性地抑制前梨状皮层神经元的延迟期活动显著地降低了工作记忆任务的行为准确率。电生理记录表明前梨状皮层神经元在感觉输入、延迟、抉择期都有对嗅觉信息的编码。（3）整体自发活动与整体联结对认知任务的贡献。利用在四个脑区（前额叶皮层mPFC，后侧前运动区pM2，丘脑中背核MD，纹状体背中区dmCP）同时电生理记录，观察到静息时期神经元可以编码上一次的抉择与行为输出。光遗传操作自发活动可以降低 mPFC在感觉输入、延迟信息存储、抉择等时相的行为相关性。大规模网络的神经活动动态受到自发活动期间的光遗传调控，总体表现为吸引子模型中初始状态的改变。（4）发现单个神经元的高频活动可以显著性的影响局部神经环路。局部环路中21%的神经元膜电位显著变化，10%的神经元动作电位发放频率发生显著变化，而整体神经元的发放升高。在细胞中进行去极化和超极化电流的注入，撤销去极化/超极化电流对后续的神经元发放并没有显著的改变。（5）发现背侧无颗粒前岛叶（AICd）与mPFC存在双向联结，其延迟期活动被抑制后，嗅觉工作记忆的行为正确率下降；光遗传调控mPFC到AICd的投射，可以双向调控工作记忆行为正确率。AICd脑区在延迟期编码嗅觉信息；抑制了mPFC到AICd的投射后，AICd群体对嗅觉信息的编码能力下降。

内容获取失败，请点击重试