RESUMEN
BACKGROUND: Golgi stains are notoriously capricious, particularly when applied to human brain. The well-known difficulties, which include complete failure of impregnation, patchy staining, unstable staining, and extensive crystalline deposits in superficial sections, have discouraged many from attempting to use these techniques. A reliable method that produces uniform impregnation in tissue from human autopsies and experimental animals is needed. NEW METHOD: The method described, "NeoGolgi", modifies previous Golgi-Cox protocols (Glaser and Van der Loos, 1981). Changes include: much longer time (>10 weeks) in Golgi solution, agitation on a slowly rocking platform, more gradual infiltration with Parlodion, more thorough removal of excess staining solution during embedding, and shorter exposure to ammonia after infiltration. RESULTS: The procedure has successfully stained over 220 consecutive frontal or hippocampal blocks from more than 175 consecutive human autopsy cases. Dendritic spines are easily recognized, and background is clear, allowing examination of very thick (200 µm) sections. Stained neurons are evenly distributed within cortical regions. The stain is stable for at least eight years. Most importantly, all stained neurons are apparently well-impregnated, eliminating ambiguity between pathology and poor impregnation that is inherent to other methods. COMPARISON WITH EXISTING METHODS: Most methods of Golgi staining are poorly predictable. They often fail completely, staining is patchy, and abnormal morphology is often indistinguishable from poor impregnation. "NeoGolgi" overcomes these problems. CONCLUSION: Starting with unfixed tissue, it is possible to obtain Golgi staining of predictably high quality in brains from human autopsies and experimental animals.