Masaki Nomura1,2*, Daisuke Ito3, Hiroki Tamate3, Kazutoshi Gohara3 and Toshio Aoyagi1,2
1CREST, JST, Kawaguchi, Saitama, Japan
2Kyoto University, Yoshida-honmachi, Sakyo-ku, Kyoto, Japan
3Hokkaido University, Kita-ku, Sapporo, Japan
*E-mail address: email@example.com
(Received November 30, 2008; Accepted February 3, 2009)
Abstract. Rat cortical neurons were cultured, and the multi-dimensional activities from the culture were recorded using 8 × 8 multi-electrode arrays. As the neurons grew, they built rich synaptic connections, and burst-like population activities were observed in the culture. Using the data during burst-like population activities, the functional connectivity between electrodes was estimated by a dynamic Bayesian analysis. A connection matrix and intrinsic firing rates were obtained from estimation. Then, the binary neuronal network model was simulated with the estimated connection matrix and intrinsic firing rates. The model was found to capture the burst-like population activities. Furthermore, the effect of excitation and inhibition balance on burst-like population activities was explored.
Keywords: Network Bursting, Functional Connectivity, Hierarchical Dynamic Bayesian Network, MCMC