Abstract: Single ion channel current signal of cell membrane is a stochastic ionic current in the order of picoampere（PA）.The background noise always dominates in the patch-clamp recordings.At present，the threshold detection method is used to restore channel current signal.However，the current threshold need be setted artificially，and this method cannot work satisfactorily when signal-to-noise ratio is lower.Hidden Markov Model（HMM） is adopted for restoring the ion-channel current and estimating parameters of the model.HMM on ion-channel is described and analyzed.Iterative algorithm based Baum-Welch is used for training HMM and estimating model parameters.Viterbi algorithm is adopted for restoring the best state sequence of ion-channel currents.Comparing HMM with the threshold detection method，the algorithm based HMM is simulated under the different transition probability and signal-to-noise ratio.The experimental results have shown that compared with the threshold detector，HMM has strong ability of anti-noise，high restoration precision，and fast convergent rate under the low signal-to-
noise ratio.Moreover，the restored error appears mainly on the mutational points of the current signal.

Key words: ionic single-channel, threshold detector method, Hidden Markov Model（HMM）, current restoration

摘要： 细胞膜离子单通道电流十分微弱（PA级），用膜片钳技术测量离子电流往往淹没在强噪声背景中。目前，采用阈值检测方法恢复通道电流信号。但是，通道开放和关闭的电流阈值需要人为设定，并且阈值法在较低信噪比时失效。采用隐马尔可夫模型（HMM）重构离子单通道电流并估计模型参数。对离子通道HMM进行描述和分析;运用Baum-Welch迭代算法训练HMM并估计模型参数;利用Viterbi算法重构通道电流最佳状态序列。将HMM与阈值法进行比较，对不同信噪比和不同转移概率情况下HMM恢复算法进行计算机仿真。结果表明：HMM与阈值法相比，具有较强抗噪能力。在较低信噪比情况下，该模型恢复信号精度高，参数收敛速度快，且电流重构误差主要出现在状态突变点。

关键词: 离子单通道, 阈值检测法, 隐马尔可夫模型, 电流重构