Distributed Synchrony - understanding the Brain at the Single Cell Level
How does the brain process information? Can it be understood at the single cell level? This paper proposes a predictive coding model based on the use of distributed synchronous spikes. We test the hypothesis that information is encoded in neurons probability of firing, which are updated spatial-temporally at precise moments. Therefore, synchronous spikes are distributed across groups of cells in a time-varying fashion. The dynamic nature of synchronous oscillation may account for the relative difficulty of detecting precisely timed spikes in experiments as well as the irregular firing patterns of single neurons. Through the use of distributed synchronous spikes, we show that a sparse, overcomplete representation of information can be learned under the governance of a predictive coding principle. Neurons in this network develop localized and oriented receptive fields, and their spike trains appear random.
Zuohua Zhang Dana H. Ballard
Department of Computer Science University of Rochester Rochester, NY 14627 USA
国际会议
8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)
上海
英文
954-959
2001-11-14（万方平台首次上网日期，不代表论文的发表时间）