Multiunit, multielectrode spike sorting software

多单元、多电极尖峰分选软件

• 批准号: 6736931
• 负责人: CHARLOTTE Mason GRUNER
• 金额: $ 50.59万
• 依托单位: R.C. ELECTRONICS, INC.
• 依托单位国家: 美国
• 财政年份: 2001
• 资助国家: 美国
• 起止时间: 2001-04-14 至 2005-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/6736931
• 关键词: action potentials; computer data analysis; computer program /software; computer system design /evaluation; electrodes; electrophysiology; mathematics; neurons; neurophysiology; time resolved data

DESCRIPTION (provided by applicant): Multiple extracellular electrodes can be used in concert to record the activity of a neural population, elucidating the complex process by which neurons code, process, learn, and retain information. Because evidence suggests that information coding involves coordinated activity of groups of neurons, such simultaneous recordings of whole neural populations are necessary for understanding coding. However, in the complex data records that result, individual neural signals are often difficult to distinguish. As a result, most researchers who collect multielectrode data must also develop their own data-processing software to determine the time-of-occurrence of action potentials, or "spikes" from the recorded neural population, a process known as "spike-sorting." The objective of this project is to develop MATLAB software toolbox for multichannel, multielectrode spike sorting that consolidates a variety of spike-sorting solutions and that incorporates state-of-the-art algorithms using advanced signal processing techniques. This will be the first commercial software package to offer effective solutions to two major spike-sorting difficulties: discrimination and identification of spike superpositions, and quantification of spike-sorting accuracy. In addition, the product will be open-source to allow users to easily add or modify modules and promote algorithm comparisons and sharing between investigators.

描述（由申请人提供）：多个细胞外电极可协同使用以记录神经群体的活动，阐明神经元编码、处理、学习和保留信息的复杂过程。因为有证据表明，信息编码涉及神经元群的协调活动，所以对整个神经群体的同步记录对于理解编码是必要的。然而，在产生的复杂数据记录中，单个神经信号通常难以区分。因此，大多数收集多电极数据的研究人员还必须开发自己的数据处理软件，以确定动作电位发生的时间，或从记录的神经群体中确定“尖峰”，这一过程被称为“尖峰分选”。" 该项目的目标是开发MATLAB软件工具箱，用于多通道，多电极尖峰分选，整合各种尖峰分选解决方案，并采用先进的信号处理技术，采用最先进的算法。这将是第一个商业软件包，提供有效的解决方案，以两个主要的尖峰分选困难：歧视和识别尖峰叠加，量化尖峰分选精度。此外，该产品将是开源的，允许用户轻松添加或修改模块，并促进研究者之间的算法比较和共享。