在大腦中，神經元可以用發放來傳遞資訊，並可用發放的時間點來增加資訊的含量。然而，至今尚無法完全了解對於神經元精準的發放的機制。過往研究指出，給予神經元一隨時間波動的刺激能使神經元產生相似性高與準確性高的發放模式(firing pattern)。本實驗研究有兩個部分：刺激波動的快慢與發放時間的影響，另一個為使用一額外控制迴路來控制神經元到目標時間時發放。首先，我們使用膜片鉗技術給予不同相關時間(τ)的Gaussian Exponential Correlated Current﹑得到不同快慢的波動來刺激神經元。細胞樣品使用初代皮質細胞培養的神經元細胞，並在9~13DIV之間進行量測。結果顯示在輸入波動的相關時間為2~10ms時，其發放模式有很高的可靠性和時間準確性。我們另外使用斜波實驗探討刺激上升斜率與發放時間之關係，結果顯示發放時間的準確率與上升斜率成正相關。最後控制實驗中，我們使用隨機最佳控制(Stochastic optimal control)的數值方法，結果顯示成功的在真實的神經元中，使發放時間與目標時間差維持在幾個毫秒內(1~5ms)。;Precise neural firing is important in the transferring of information. However, we still do not fully understand the mechanism that allows neurons to generate a spike within a narrow temporal regime in the order of a few millisecond. To study the spike timing in a neuron, we perform two types of experiment. Firstly, we stimulate the neuron with different temporal correlated stimuli using patch clamp, and analyze the statistics of corresponding firing patterns; secondly, we apply the stochastic optimal control to manipulate the spiking time. Our samples are from the primary cortical culture. Our result indicates that precision of spike timing is optimal when the correlation times of stimuli are from 2 to 10 ms. In the ramp stimuli experiment, the higher slope of ramp stimuli result in higher precision of spike timing. Obviously, the precision of spike timing depends on stimulus transients. In the stochastic optimal control experiment, the spike timing can be controlled within a few millisecond(1~5ms).
