Pitfalls of using confocal-microscopy based automated quantification of synaptic complexes in honeybee mushroom bodies (response to Peng and Yang 2016)

A recent study by Peng and Yang in Scientific Reports using confocal-microscopy based automated quantification of anti-synapsin labeled microglomeruli in the mushroom bodies of honeybee brains reports potentially incorrect numbers of microglomerular densities. Whereas several previous studies using visually supervised or automated counts from confocal images and analyses of serial 3D electron-microscopy data reported consistent numbers of synaptic complexes per volume, Peng and Yang revealed extremely low numbers differing by a factor of 18 or more from those obtained in visually supervised counts, and by a factor 22–180 from numbers in two other studies using automated counts. This extreme discrepancy is especially disturbing as close comparison of raw confocal images of anti-synapsin labeled whole-mount brain preparations are highly similar across these studies. We conclude that these discrepancies may reside in potential misapplication of confocal imaging followed by erroneous use of automated image analysis software. Consequently, the reported microglomerular densities during maturation and after manipulation by insecticides require validation by application of appropriate confocal imaging methods and analyses tools that rely on skilled observers. We suggest several improvements towards more reliable or standardized automated or semi-automated synapse counts in whole mount preparations of insect brains.