@misc{oai:ir.soken.ac.jp:00002477,
 author = {中江, 健 and ナカエ　, ケン and NAKAE, Ken},
 month = {2016-02-17, 2016-02-17},
 note = {Phase response curve (PRC) describes the response of an oscillator to external perturbation; it is useful to predict and understand synchronized dynamics of oscillators. In recent years, neuroscientists have focused on neurons’ PRCs, and measured them experimentally. This originates from the leading hypotheses that the synchronization of neurons has a functional meaning in the brain.<br/>
In this thesis, we propose two statistical methods for estimating PRCs from data; it allows for the correlation of errors in explanatory and response variables of the PRC. This correlation is neglected in previous studies.<br/>
The ?rst method is implemented with a replica exchange Monte Carlo technique; this avoids local minima and enables ef?cient calculation of posterior averages. A test with arti?cial data generated by noisy Morris-Lecar equations shows that, in terms of accuracy, this method outperforms conventional regression that ignores errors in the explanatory variable. Experimental data from the pyramidal cells in the rat motor cortex is also analyzed; a case is found where the result with the ?rst method is considerably different from that obtained by conventional regression.<br/>
The second method is developed using a transformation that mixes the variables; it effectively removes the correlation. The computation time of this method is considerably less than that of the ?rst method. The same test using the noisy Morris-Lecar equations shows that the second method also outperforms than convectional regression in terms of accuracy.<br/>, application/pdf, 総研大甲第1420号},
 title = {Statistical Estimation of Phase Response Curves},
 year = {}
}