ウイナー非線形解析:神経情報処理研究における応用
スポンサーリンク
概要
- 論文の詳細を見る
Physiological systems can be identified systematically and efficiently by use of the nonlinear or white-noise analysis technique whose theoretical ground was laid by Wiener in 1942. The technique uses statistical, unbiased random noise as input to probe a system and identifies it through the process of cross-correlation, whereas the traditional approach uses purpose-oriented inputs to prove or disprove a particular and often limited aspect of a systems function.<BR>Since our first application of the technique to define transfer functionals of neuron chains in catfish retina, many advances have been made in the theoretical as well as practical aspects of the theory. An introductory book on the subject was written by the Marmarelis brothers (Marmerelis and Marmarelis, 1977) and the latest theoretical advance is seen in Yasuis paper (Yasui, 1979). White-noise signals to probe biological systems now encompass such inputs as spatio-temporal random light signals, current injected into a point in a complex nervous system, and binary or ternary signals.<BR>In our study of catfish retina, (first order) Wiener kemels were related to such traditional but moot concepts as the Weber-Fechner relationship or the Michaelis-Menten equation. Temporal as well as spatial dynamics of neurons in catfish retina have been defined by use of multi-input random signals or the one-dimensional random travelling grating devised by Yasui.<BR>Although the technique may not be immediately accepted by biologists, it will, in the long run prove to be a valuable tool for studying complex biological systems.
- 日本生物物理学会の論文