Distortions induced in neuronal quantification by camera lucida analysis: comparisons using a semi-automated data acquisition system.

Quantitative analysis of dendritic structure is widely used in assessing neural relationships in Golgi-stained tissue. Quantitative techniques are tedious and time-consuming. A computer-microscope system is described which speeds data acquisition and analysis. Three neuronal samples are analyzed both by this computerized system and by camera lucida techniques. We demonstrate that the computerized techniques result in far fewer errors in data transcription and analysis than the camera lucida procedure. In addition, we show that the amount of distortion in camera lucida drawings caused by collapsing 3 dimensions into 2 varies both with cell class and cell dendrite type. This prevents the use of any simple statistical procedures for deriving 3-dimensional information from 2-dimensional data.