This layer was defined as the mitral cell layer (MCL). In some preparations, mitral cells were confirmed by histology (data not shown). The layer that was intermediate to the GL and MCL was defined as the external plexiform layer (EPL; depth 100–250 μm). The depths of individual neurons

were normalized to the depths of the mitral cell layer for each sample, with, the brain surface defined as 0.0 and the MCL as 1.0. Based on the cell layers, cell sizes, cell shapes, and the presence or absence of L-Dends, labeled neurons were categorized into six neuronal subtypes (Table S1). Three cell subtypes were distinguished in the GL based on morphological structures (Figure 1 and Table S1). Small cells (n = 30) did not have L-Dends and were assumed to be periglomerular Selleck RG-7204 cells. The middle cells with/without L-Dends were considered

to be external tufted cells (n = 53 and n = 37, respectively). A portion of the anatomically identified neurons was used for functional analysis. Because significant differences in eMRR widths and similarities could not be detected P450 inhibitor in the GL, the data from these three cell subtypes in the GL were combined and referred to as juxtaglomerular (JG) cells for functional comparisons. In the EPL, two types of cells were distinguished: cells with L-Dends and cells without L-Dends. The majority of these projection neurons in the EPL were considered to be middle tufted cells. For unclear reasons, odor-induced Ca2+ responses were only successfully recorded from cells with L-Dends. In the MCL, all of the labeled MCL cells had L-Dends. The majority of projection neurons in the MCL were considered to be mitral cells. Using heterozygous OMP-Synapto-pHluorin knockin mice (Bozza et al., 2004), olfactory sensory axon terminal glomerular activities were detected using a microscope (BX50WI; Olympus) that was equipped with a high speed CCD camera (NeuroCCD-SM256; Redshirt Imaging). The OB was illuminated with an LED light at 470 nm (M470L2, Thorlab). The excitation and emission lights were band-pass filtered with a GFP

filter set (BrightLine Fossariinae GFP-4050A, Semrock) and collected at 25 Hz. Raw fluorescence traces from individual glomeruli were sampled by spatial averaging of 3–4 pixels that were located near the center of each glomerulus. Photobleaching was corrected by subtracting fluorescent responses observed during a no-odor imaging trial. Each series of images was evaluated by subtracting the resting fluorescence (F) (average of 75 images for 3 s prior to odorant delivery) and the resulting values were expressed as ΔF/F. Mann-Whitney tests were used to determine significant odor-evoked responses by comparing the averaged images before (for 3 s) and after odor onset (for 6 s). Differences of p < 0.05 were considered to be statistically significant.