INFORMATION THEORETIC ANALYSIS OF NEURAL CODES

描述(改编自申请者的摘要)： 拟议的研究是确定关于感觉刺激的信息如何 特征由集合的神经活动模式编码。准备工作 要研究的是板球的颈椎感觉系统。具体的 这一提议的目的是检验以下假设：1)动态 使用时间编码方案来表示感觉刺激的特征； 2)使用集合编码方案来表示感觉刺激；3) 没有重大的行为相关信息的损失 传入细胞阵列和20细胞系综间的突触接口 初级感觉中间神经元；以及，4)感觉系统被优化以 处理来自动物环境的典型行为相关信号。 研究的实验数据库将通过ADVERCESS 电生理记录技术。技术将包括DUAL 神经元对和细胞外多单位的细胞内记录 紧张的录音。将实现细胞外录制 通过使用专门设计的线和硅记录阵列 用于这些研究。分析方法将使用信息论， 随机系统分析和几个工程分析。神经网络 算法可以用来实现复杂结构的非线性回归分析 数据集。

DESCRIPTION (Adapted from applicant's abstract): The general goal of the proposed research is to determine how information about sensory stimulus features is encoded by ensemble neural activity patterns. The preparation to be studied is the cercal sensory system of the cricket. The specific aims of this proposal are to test the following hypotheses: 1) A dynamic temporal encoding scheme is used to represent features of sensory stimuli; 2) A collective encoding scheme is used to represent sensory stimuli; 3) There is no significant loss of behaviorally relevant information across the synaptic interface between the afferent array and the 20 cell ensemble of primary sensory interneurons; and, 4) The sensory system is optimized to process typical behaviorally relevant signals from the animal's environment. The experimental database for the studies will be collected through advanced electrophysiological recording techniques. Techniques will include dual intracellular recordings from pairs of neurons and multi-unit extracellular recordings from nerves. The extracellular recordings will be achieved through the use of wire and silicon recording arrays designed specifically for these studies. The analytical approach will employ information theory, stochastic systems analysis and several engineering analyses. Neural net algorithms may be used be implement nonlinear regression analyses of complex data sets.