Quick Search:
Author: Title/Abstract: Vol./No: Page:

Prog. Theor. Phys. Supplement No.157 (2005) pp. 300-303

[ Full Text PDF : FREE ACCESS (112K) ]

Statistical Mechanics for Neural Spike Data Analysis Using Log-Linear Model

Hayaru Shouno,1 Koji Wada2 and Masato Okada3,4,5

1Faculty of Engineering, Yamaguchi University, Ube 755-8611, Japan
2Kochi National College of Technology, Nangoku 783-8508, Japan
3RIKEN BSI, Wako 351-0198, Japan
4Japan Scientific Technology Corp., Kashiwa 277-8561, Japan
5Graduate School of Frontier Science, The University of Tokyo, Kashiwa 277-8561, Japan

Abstract:

Recently, we can simultaneously record spike data from many neurons in the field of electrophysiology, and thus it is required to develop mathematical framework for extracting higher-order correlation of neural firings. The joint probability of neural spike can be represented using the log-linear model. From statistical-mechanical point of view, the log-linear model can be regarded as a multi-body interacted Ising spin model or the Boltzman machine with higher-order interactions. The estimation of higher-order correlation of neural firing corresponds to that of higher-order interations Ising spin system, and to the hyper-parameter estimation in the Bayesian inference. In this paper, we apply maximization of marginal likelihood (MML) method to this problem, and discuss the properties of MML analytically using statistical-mechanical method.


URL : http://ptp.ipap.jp/link?PTPS/157/300/
DOI : 10.1143/PTPS.157.300

[ Full Text PDF : FREE ACCESS (112K) ] Citation:


References:

  1. H. Nakahara and S. Amari, Neural Computation 14 (2002), 2269.
  2. K. Wada and M. Okada, The IEICE Trans. Inf. Syst. 85-D-II (2002), 1582.
  3. J. Inoue and K. Tanaka, Phys. Rev. E 65 (2002), 016125[APS].
  4. J. Inoue and K. Tanaka, J. of Phys. A 36 (2003), 10997[IoP STACKS].
  5. A. P. Dempster, N. M. Larid and D. B. Rubin, J. Roy. Statist. Soc. B 39 (1977), 1.