Specificity of spatial patterns of glomerular activation in the mouse olfactory bulb: computer-assisted image analysis of 2-deoxyglucose autoradiograms.

We have developed a computer-assisted method for analyzing the 2-deoxyglucose (2-DG) autoradiograms of mice olfactory bulbs. The purpose of the study was to numerize the maps of glomerular activation in order to achieve a statistical comparison of the glomerular patterns evoked by different stimuli. The spatial distribution of glomerular activation was displayed on unfolded representations of the glomerular layer which were built up using glomerular optical densities (OD) measured systematically within 13 sections per bulb. Each bulbar sample was converted into an 'OD profile'. A matrix composed of 18 OD profiles was submitted to a principal component analysis. The first factor which accounted for 28% of the variance separated unambiguously two clusters corresponding to the bulbs issued from animals stimulated with amylacetate and isovaleric acid, respectively. The second and third factors which accounted for 14% and 12% of the variance segregated the control group (animals exposed to pure air) from the odor-stimulated ones. It was demonstrated that the cluster separation was actually due to the specific spatial distribution of the most-labelled glomeruli. A particular attention was paid to the well-delineated glomerular activation evoked by isovaleric acid. The results demonstrate the specificity and reliability of the glomerular 2-DG patterns. The method should be useful for further comparisons of patterns elicited by larger sets of odorant compounds.