LOCAL OPTIMIZATION OF NEURON ARBORS

被引：72

作者：

CHERNIAK, C

机构：

[1] Department of Philosophy, University of Maryland, College Park, 20742, MD

来源：

BIOLOGICAL CYBERNETICS | 1992年 / 66卷 / 06期

关键词：

D O I：

10.1007/BF00204115

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

How parsimoniously is brain wiring laid out, that is, how well does a neuron minimize costs of connections among its synapses? Neural optimization of dentritic and axonic arbors can be evaluated using a generalization of the Steiner tree concept from combinatorial network optimization theory. Local branch-junction geometry of neuronal connecting structures fits a volume minimization model well. In addition, volume of the arborizations at this neighborhood level is significantly more strongly minimized than their length, signal propagation speed, or surface area. The mechanism of this local volume optimization resembles those involved in formation of nonliving tree structures such as river junctions and electric-discharge patterns, and appears to govern initial nerve growth-cone behavior through vector-mechanical energy minimization.

引用

页码：503 / 510

页数：8

共 42 条

[1]

BANKER G, 1988, INTRINSIC DETERMINAN

[2] THE SHORTEST-NETWORK PROBLEM [J].