Miller A K
J Neurosci Methods. 1981 Dec;4(4):407-30. doi: 10.1016/0165-0270(81)90009-1.
An approach to the automatic counting of neurones in the human hippocampus is described which uses a modified Quantimet 720 image analyzing computer. Previous work with this and similar machines has attempted to use size alone to discriminate between the various types of cell which are present in any section of central nervous system tissue. Serious errors may then result since agglomerates of small cells and cell fragments can bae confused with single large cells. Instead, 2 electron devices have been designed and built which allow nucleolated neurones, defined as those objects having a grey phase and a black phase both within designated area limits, to be distinguished from all other cells. The mean values for absolute cell numbers and for cell density were within 5% of those obtained by traditional manual methods. The operation of both electronic devices is described and circuit diagrams given for one of them. The other is beyond the scope of this paper, but has been fully reported elsewhere.
本文描述了一种使用改进型Quantimet 720图像分析计算机自动计数人类海马体神经元的方法。此前使用该机器及类似设备的工作试图仅依据大小来区分中枢神经系统组织任何切片中存在的各种类型的细胞。然而,这样可能会导致严重错误,因为小细胞聚集体和细胞碎片可能会与单个大细胞混淆。取而代之的是,已经设计并制造了两种电子设备,它们能够将核仁神经元(定义为在指定面积范围内同时具有灰色相和黑色相的物体)与所有其他细胞区分开来。绝对细胞数量和细胞密度的平均值与通过传统手工方法获得的数值相差在5%以内。文中描述了这两种电子设备的操作,并给出了其中一种的电路图。另一种设备超出了本文的范围,但已在其他地方进行了全面报道。