Estimating amperometric spike parameters resulting from quantal exocytosis using curve fitting seeded by a matched-filter algorithm.
Estimating amperometric spike parameters resulting from quantal exocytosis using curve fitting seeded by a matched-filter algorithm.
复制标题
使用由匹配滤波器算法播种的曲线拟合来估计由量子胞吐作用产生的电流尖峰参数。
DOI：
10.1016/j.jneumeth.2018.09.018
复制
发表时间：
2019
影响因子：
3
通讯作者：
Gillis,KevinD
中科院分区：
文献类型：
--
作者：
BalajiRamachandran,Supriya;Gillis,KevinD
背景使用电化学微电极可以检测到可氧化神经递质的量子胞吐作为安培电流的尖峰。对棘波参数的测量表明了最大的发射器流量、流量持续时间和从单个囊泡释放的发射器的量。自动分析算法需要拒绝在时间上重叠的峰值。此外，许多棘波之前都有小幅度的“脚”信号，归因于递质通过融合孔的缓慢释放。准确的尖峰前基线确定对于估计融合孔持续时间和通过融合孔释放的递质量至关重要。新方法我们开发了一种基于多指数函数对数据进行拟合的估计方法。我们前面描述的匹配过滤器算法被用来识别要拟合的数据段，并提供种子值以促进迭代拟合的收敛。新的估计算法包括重叠剔除、两步拟合法和一种新的基线估计方法。结果SPEKE参数的历史曲线表明新方法与人工计算的参数具有很好的一致性。与现有方法相比，使用新方法生成的参数估计比常用的现有方法更接近于盲人手估计。性能的提高是由于对有效尖峰的更好检测和对重叠尖峰的拒绝。结论匹配滤波种子算法能够可靠地抑制重叠，以全自动的方式估计棘波和足部信号参数。
BackgroundQuantal exocytosis of oxidizable neurotransmitters can be detected as spikes of amperometric current using electrochemical microelectrodes. Measurements of spike parameters indicate the maximal transmitter flux, flux duration, and amount of transmitter released from individual vesicles. Automated analysis algorithms need to reject spikes that overlap in time. In addition, many spikes are preceded by small amplitude “foot” signals, attributed to slow release of transmitter through a fusion pore. Accurate pre-spike baseline determination is essential for estimating fusion-pore duration and the amount of transmitter released through the fusion pore.New MethodWe developed an estimation approach that is based on fitting a multi-exponential function to the data. Our previously described matched-filter algorithm is used to identify the sections of data to fit and provides seed values to facilitate convergence of the iterative fit. The new estimation algorithm includes overlap rejection, a two-step fitting procedure and a novel baseline estimation procedure.ResultsHistograms of spike parameters demonstrate excellent agreement of the new approach with manually computed parameters.Comparison with Existing MethodsParameter estimates generated using the new approach are closer to blind manual estimates than commonly used existing methods. The improved performance is due to better detection of valid spikes and rejection of overlapping spikes. Moreover, since the complete time course of the spike is fit to a function, more complete information about the spike time course is captured.ConclusionsThe matched-filter seeded algorithm reliably rejects overlaps and estimates spike and foot signal parameters in a fully automated manner.