Identification at Parallel Distributed Processing in the Plasticity of Neural Networks by Using on Optical Recording System

1. 我们研制了144通道光学记录系统，成功地记录了海马神经元的多单元活动。利用该方法鉴定了CA1神经元在时间模式刺激下的不同LTP。我们发现，即使在刺激集的平均频率保持不变的情况下，增强不仅取决于刺激的频率，而且取决于连续间隔之间的相关程度。这表明刺激的详细时间特性是诱发LTP的重要因素，因此可能在学习和记忆现象中起重要作用。基于上述生理数据，我们提出了海马CA1区LTP诱导的数学模型。模型学习规则本质上是Hebbian性质的，即LTP的诱导量随着突触激活频率的增加而增加，但它对应用刺激链的刺激间隔之间的相关性也有敏感性。基于生理证据，我们提出了事件记忆的关联和分离是如何在海马-皮层记忆系统中形成的假说。据推测，记忆的关联/分离是通过索引事件（记忆内容）的表征与时间代码来进行的。

1. We developed a 144-channels optical recording system, by which we succeeded in recording multi-units activities of hippocampal neurons. By using the method, the different LTP of CA1 neurons induced by the temporal pattern stimuli were identified.2. We find that the potentiation depend not only on the frequency of the stimulus, but on the degree of correlation between successive intervals, even when the average frequency of the stimulus set is held constant. This suggests that detailed temporal properties of the stimulus are important factors for inducing LTP, and hence could play a crucial role in learning and memory phenomenan.3. We proposed a mathematical model for the induction of LTP in the CA1 region of the hippocampus based on the physiological data obtained above. The model learning rule is essentially Hebbian in nature in the sense that the amount of the LTP induced increases with frequency of synaptic activation, but it also has a sensitivity on the correlation between inter-stimulus intervals of applied stimulus chain.4. Based on the physiological evidence, we proposed a hypothesis on how association and dissociation of event memories are made in the hippocampal-cortical memory system. It is postulated that association /dissociation of memory is carried out by indexing the representations of events (memory contents) with temporal codes.