The dynamic connectome: dynamics of learning

大脑皮层的连接组是高度动态的，即使在基础条件下也表现出突触连接的高周转率。然而，我们的大脑能够保持终身的记忆。在这样一个动态的环境中，这些记忆是如何形成和保护的？在这里，我们建议将啮齿动物皮层在学习过程中兴奋性和抑制性突触连通性的延时成像与高通量自动数据分析和计算建模相结合，以帮助回答这个基本问题。我们在第一个资助期的合作努力为这项任务奠定了重要的方法和分析基础：首先，我们将使用在第一个资助期开发的时间推移成像技术来测量小鼠在学习听觉提示go/nogo任务之前，期间和之后听觉皮层的兴奋性和抑制性连通性的动态。其次，我们将应用并进一步完善基于深度神经网络的高通量自动图像分析技术，以执行兴奋性和抑制性连接动态的自动量化。第三，我们将使用计算建模来描述和解释在学习过程中观察到的连接体动态，以揭示可能的潜在机制，并为未来的实验生成可测试的预测。结合三个实验室的互补专业知识，我们的目标是将目前对学习诱导的突触可塑性的描述从单树突或单神经元扩展到连接组水平。

The connectome of the cerebral cortex is highly dynamic, exhibiting high turnover of synaptic connections even under basal conditions. Nevertheless, our brains are able to maintain life-long memories. How are such memories formed and safeguarded in such a dynamic environment? Here, we propose to combine time lapse imaging of excitatory and inhibitory synaptic connectivity of rodent cortex during learning with high-throughput automated data analysis and computational modeling to help answer this fundamental question. Our collaborative effort during the first funding period has laid the crucial methodological and analytical foundation for this task: First, we will use time lapse imaging technology developed during the first period of funding to measure the dynamics of excitatory and inhibitory connectivity in the auditory cortex of mice before, during, and after learning of an auditory cued go/nogo task. Second, we will apply and further refine high throughput automated image analysis techniques based on deep neural networks developed during the first funding period to perform automated quantification of the dynamics of excitatory and inhibitory connections. Third, we will use computational modeling to describe and explain the observed connectome dynamics during learning to reveal possible underlying mechanisms and generate testable predictions for future experiments. Combining the complementary expertise from three laboratories, we aim to extend current descriptions of learning-induced synaptic plasticity from the single-dendrite or single-neuron towards the connectome level.