Bukh-Viner P V, Feshchenko V A, Chilingarian L I
Zh Vyssh Nerv Deiat Im I P Pavlova. 1992 May-Jun;42(3):609-14.
The collector is an adaptive algorithm for pattern recognition. It proposes new in-line fully-automatic technique to learn and recognize effective patterns of input data stream. Evoked potentials (EP) were recorded by ADDA 100 KHz, 4 channels, and described by 200 points per each EP. The collector recognized different studies of conditioned response (CR) by patterns of EPs in amygdalar central nucleus. In dogs with implanted into the limbic structures concentric electrodes an instrumental CR was elaborated to electrical stimulation of the dorsal hippocampus. Generalization or transfer of this CR was tested by means of electrostimulation of amygdalar basal nucleus. The generalization in the first experiment took place approximately in 86% of cases, in the second one in 52% of cases. In the first experiment the amplitudes of initial negativity and of late positive waves were smaller than those in the second one and in the experiments before conditioning.
收集器是一种用于模式识别的自适应算法。它提出了新的在线全自动技术来学习和识别输入数据流的有效模式。诱发电位(EP)由ADDA以100千赫兹、4通道记录,每个EP由200个点描述。收集器通过杏仁核中央核中EP的模式识别不同的条件反应(CR)研究。在植入边缘结构同心电极的狗中,对背侧海马体的电刺激形成了工具性CR。通过杏仁核基底核的电刺激测试了这种CR的泛化或转移。在第一个实验中,约86%的情况发生了泛化,在第二个实验中为52%。在第一个实验中,初始负波和晚期正波的振幅小于第二个实验以及条件反射前实验中的振幅。