Odor Information Representation Using the Temporal Dimension : Concentration-Invariance and Speed-Accuracy Trade-off.

动物有很强的感知各种气味的能力，即使环境是动态和不可预测的变化。在这里，我们提出了一个计算模型，其中的浓度不变的气味信息表示的时空活动的网络。所表示的气味信息很容易评估和存储的另一个网络模型的更高的嗅觉中心。在我们的模型中，将输入空间模式转换为不变的时空顺序是表示气味信息的关键，并且所提出的网络的输出模式与生物网络的输出模式一致。

Animals have an excellent ability to perceive various odors robustly, even when the environment is dynamically and unpredictably changing. Here, we proposed a computational model in which the concentration-invariant odor information is represented by the spatiotemporal activity of network. The represented odor information is easily evaluated and stored by another network model for the higher olfactory center. In our model, transforming the input spatial pattern inlo the invariant spatiotemporal order is essential for representing the odor information and output patterns of the Proposed networks are consistent with those of the biological ones.