摘要
本文提出了一种简化现实性神经网络模型的离散形式，特别适合于研究其复杂行为．适当联结的各神经元分级电位的某种总和可表现出类似于脑电的复发行为，虽然其所受到的刺激全部是相同的恒定刺激，也没有噪声．周期性地迭加上方波刺激然后将相应的反应进行迭加可以诱发出类似于在核团附近记录的双相反应．
ased on a simplifying realistic neural network model proposed previously by one of the authors, a discrete form of the model is developed in this paper, which is especially suited for studying complex behaviour. Some summation of the graded potentials of all the neurons with proper interconnection could manifest complex behaviour similar to EEG, when the stimuli are all the same constant ones and with no noise. Superimposing square-wave stimuli periodically and averaging the data could 'evoke'biphasic response somewhat similar to EP recorded near nucleus.
出处
《复旦学报（自然科学版）》
CAS
CSCD
北大核心
1996年第4期447-451,共5页
Journal of Fudan University：Natural Science
基金
国家攀登计划和国家自然科学基金
