Summary
International Symposium on Nonlinear Theory and its Applications
2010
Session Number:A2L-D
Session:
Number:A2L-D2
Estimation of Neural Network Structure by Transforming Spike Sequences to Continuous Time Series
Kaori Kuroda, Tohru Ikeguchi,
pp.123-126
Publication Date:2010/9/5
Online ISSN:2188-5079
Summary:
In neural systems, many complicated behaviors are observed. To understand network dynamics, and reproduce the complicated behavior, it is important to clarify the network structures as well as their dynamics. To resolve this issue, we have already proposed a measure, partial spike time metric. Although this measure exhibits high performance to estimate the network structures, it cannot evaluate negative correlations correctly. In this paper, to resolve this problem, we transform multi-spike sequences into continuous time series to estimate the neural network structures. As a result, our proposed method is more effective for estimating neural network structures than the conventional method.