神经系统在不同水平上表现出丰富的短期动力学特性，如在单个神经细胞水平上的发 放频率适应(spike frequency adaptation, SFA )，在突触水平上的短时程增强( short-term facilitation, STF )和短时程抑制( short-term depression, STD )。这些动力学特征通常覆盖广泛 的时间尺度，并且在不同的大脑区域表现出很大的多样性。目前还不清楚大脑的短期动力 学具有这种多样性的计算上的好处是什么。在本研究中，我们??出大脑可以利用这些动力 学特征具有不同的时间尺度来在单个神经环路中实现多个看似矛盾的计算。为了验证这一 思想，我们使用连续吸引子神经网络( CANN )作为工作模型，并在它的动力学中耦合时间 尺度依次增加的 STF、SFA 和 STD。我们考虑了三个计算任务，即神经细胞的持续发放、 神经细胞的适应性和对移动刺激的超前跟踪。这些任务需要相互冲突的神经机制，因此不 能通过单个动力学特征或具有类似时间尺度的任何动力学的组合来实现。然而，通过适当 地协调 STF、SFA 和 STD 的时间尺度，我们表明该网络能够同时实现这三个计算任务。我 们希望这项研究将有助于理解大脑如何在不同的层次上协调其丰富的动力学以实现不同 的认知功能。

The short-term dynamics of the nervous system show abundant levels, such as spike frequency adaptation (SFA) at single neuron levels, and short-term facilitation (STF) and depression (STD) at the synapse level. These dynamical characteristics usually cover a wide range of time scales, and show great diversity in different brain regions. It is not clear what computational benefits the brain has for short-term dynamical variability. In this study, we suggest that the brain can use this dynamical feature to achieve many seemingly contradictory calculations in a single neural circuit. To prove this idea, we use the continuous attractor neural network (CANN) as a working model and include STF, SFA and STD with increasing time constant. Three computing tasks are considered. These tasks are persistent activity, adaptation, and anticipative tracking. These tasks require conflicting neural mechanisms, so they cannot be achieved by any single dynamical property or any combination of similar time constants. However, with correctly coordinated STF, SFA and STD, we show that the network can achieve three computing tasks at the same time. We hope that this study can reveal how the brain coordinates various levels of its dynamics to achieve various cognitive functions.