Model Reference Current Injection: A High-Speed Next-Generation Dynamic Clamp

“动态钳”是近十年发展起来的一种电生理技术，它利用神经元和心肌细胞等电可兴奋细胞的实时电压测量。这些测量被实时运行的计算模型用来计算注入(相同或另一个)可兴奋细胞的人造电流。这项技术的应用包括增强单个细胞的固有电学性质，或使用计算机连接孤立的细胞，以创建由真实神经元和计算的突触组成的人工网络。这种方法在神经科学和心脏电生理学领域得到了越来越多的使用。然而，这项技术的实施并不容易，而且现成的(交钥匙)方法没有跟上快速变化的计算技术的步伐。此外，尽管许多实验室已经使用了这项技术，但很少有研究来评估它的准确性和有效性。这项应用的目标是开发一种实现动态钳位技术并利用现有计算机技术的下一代系统。这项技术将是低成本的，使用大多数神经生理学实验室中提供的计算机和电生理硬件；功能齐全，允许非常高的速度(目标为100 kHz)、任意型号规格和控制计算机上的实时数据记录；以及开源，这意味着它所基于的软件和操作系统可以自由分发和自由修改，以满足实验室的特定需求。传统的桌面操作系统不适合在10‘S千赫范围内的实时性能，因此本项目将使用实时Linux来实现这些目标，Real-time LINUX是Linux操作系统的衍生版本，经过修改以执行实时任务。该项目的具体目标是开发一个基于实时Linux的实时控制计算引擎，开发一个适用于任意模型规范的用户界面和机制，并对该技术的准确性和有效性进行严格的测试。该技术将从多个方面影响神经科学和电生理学领域的研究。首先，它将使更多的研究人员可以使用这项技术，提供更多可用的功能。其次，这种方法速度的提高将允许使用动态钳制进行新的实验，例如那些涉及具有亚毫秒时间常数的膜过程的实验，例如心肌细胞中的钠电流和来自快速AMPA突触的突触电流。

Abstract ButeraThe "dynamic-clamp" is an electrophysiological technique developed in thepast decade that utilizes real-time voltage-measurements from electricallyexcitable cells, such as neurons and cardiac myocytes. These measurementsare used by a computational model running in real-time to compute artificialionic currents that are injected into (the same or another) excitable cells. Applications of this technique include augmenting the intrinsic electricalproperties of a single cell or using the computer to connect isolated cellsto create an artificial network consisting of real neurons and computedsynapses. This approach has seen growing use in the neuroscience andcardiac electrophysiology communities. However, the technique is nontrivialto implement and off-the-shelf (turnkey) approaches have not kept pace withrapidly changing computing technologies. Furthermore, although manylaboratories have used this technique, little research has been done toevaulate its accuracy and validity.The objective of this application is to develop a next-generation systemthat implements the dynamic-clamp tecnique and utilizes currently avaiablecomputing technologies. This technique will be low-cost, using computersand electrophysiolgical hardware available in most neurophysiologicallaboratories; fully-featured, allowing for very high speeds (with a goal of100 kHz), arbitrary model specifications, and real-time data logging on thecontrol computer; and open-source, meaning that the software and operatingsystem upon which it is based are freely distributable and freely modifiableto suit a laboratory's specific needs. Conventional desktop operatingsystems are unsuitable for real-time performance in the 10's of kHz range,thus this project will use Real-Time Linux, a derivative of the Linuxoperating system modified for performing real-time tasks, to achieve thesegoals. The specific objectives of this project are to develop a corereal-time control computational engine based on Real-Time Linux, develop auser interface and mechanism for arbitrary model specification, and performrigorous testing of the accuracy and validity of the technique.This technology will impact the neuroscience and electrophysiology fields ofresearch in several ways. First, it will make the technique available tomore researchers, with a greater array of features available. Second, theincreased speed of this approach will allow new experiments to be performedusing the dynamic clamp, such as those involving membrane processes withsubmillisecond time-constants, such as sodium currents in cardiac myocytesand synaptic currents from fast AMPA synapses.