ABSTRACT
To achieve realistic simulations of biological neurons anatomically correct morphological reconstructions of the desired cell type need to be obtained first. If significant errors are made in the reconstruction, the resulting passive and active properties of the simulated neuron could substantially deviate from the real cell. For example, if the diameter of a thin dendrite was traced at 0.5
µ
m but the actual diameter was 0.8
µ
m, an error of 60% for the surface area and of 156% for the cross-sectional area would result. Of course, the surface area is proportional to the cell capacitance and total membrane conductance, and the cross-sectional area to the axial conductance, which are important parameters describing the passive structure of neural models (Chapter 8). Errors of similar magnitude easily result from omitting the surface area of dendritic spines. Other major sources of error in reconstructions include the degradation of neurons during recording, shrinkage artifacts during histological processing, the omission of faintly stained distal processes from the reconstruction, and just the limitations in the resolution of light microscopy if this technique is used without further validation. This chapter is aimed at providing a step by step description of the techniques that can be used to reduce such problems and to obtain optimal morphological reconstructions for modeling.
