【摘要】 目的 :新建一个“优势模式 (favoredpattern ,FP)”快速检测法 ,用以检测不同神经元放电串中之FP ,并检验其效果。方法 :在原有的“改进的FP检测法”的基础上 ,采用数种措施实现快速检测 ,并用多个指标检验其检测效果。结果 :(1)具有 3种不同程度变动的各种模拟FP的检出阈均为 2 ,表明该法敏感 ;(2 )自发放电串中各个FP重复出现时 ,其全长和所含不同放电间隔长度之变异系数分别为 5 %和 17% ,均小于检测所规定的容差 ,表明该法可靠 ;(3)自发放电串中FP长时间稳定地存在 ,而诱发放电串中不同FP的总数随放电率的变化而增减 ;(4 )所有被检测的放电串 (总数 92 )都能检测到FP ,检测一个 12 0 0～ 15 0 0个动作电位的放电串中的FP仅需 2～ 3s。结论 :新建立的“快速FP检测法”可能是研究神经信息编码的一个很有用的手段更多还原

【Abstract】 Objective: To develop a high speed detection method for picking up favored pattern(FP) in spike trains and to examine its efficiency. Methods: On the basis of the “modified detection method”, various procedures was used to speed up the detection and several indexes to examine its efficiency. Results: (1) Detection thresholds of simulated patterns with three different degrees of variation were 2, indicating that the method is sensitive. (2) The largest CV(coefficient of variation) of FP lengths and interval lengths in the FP repetitions were 5% and 17% respectively, less than 20% (the “tolerance of variation” used), indicating that this is reliable. (3) The number of different FPs in the long spontaneous spike train was relatively stable, whereas in evoked spike trains the number of different FPs either increased or decreased with the firing rate. (4) FPs could be identified from all spike trains detected (detection rate: 100%), and the computing time used for analyzing a spike train containing 1 200～1 500 spikes was 2～3 seconds. Conclusion: The newly developed “high speed detection method” may be a very useful tool for the study of neural coding.更多还原