A critical issue in quantitative neuromorphology is the accuracy and subsequent reliability of the tracing techniques employed to characterize neuronal components. Historically, the camera lucida was the only option for such investigations. In 1987, MBF Bioscience, Inc. (Williston, VT) developed the integrative Neurolucida computer-microscope system, replacing the camera lucida drawing tube with a Lucivid cathode ray tube, thereby allowing computer overlays directly on the view through microscope oculars. Subsequent advances in digital cameras have allowed the Lucivid system to be replaced so that microscope images can be traced by viewing the digital image on a computer monitor. Indeed, the camera systems now outsell Lucivid systems 9 to 1 (J. Glaser, personal communication, 08/2008). Nevertheless, researchers seldom note which of these configurations are being used (which may confound the accuracy of data sharing), and there have been no published comparisons of the Lucivid and camera configurations. The present study thus assesses the relative accuracy of these two hardware configurations by examining reconstructions of human pyramidal neurons. We report significant differences with respect to dendritic spines, with the camera estimates of spine counts being greater than those obtained with the Lucivid system. Potential underlying reasons (e.g., magnification, illumination, and resolution, as well as observer ergonomic differences between the two systems) for these quantitative findings are explored here, along with qualitative observations on the relative strengths of each configuration.
Read full abstract