Epithelial chromatography may be related to chemotopic progressions in the olfactory bulb. A, D: The structure of the nasal cavity is shown as a diagram of a coronal section taken just anterior to the caudal end of the internal naris. Odorant representations in the medial aspect of the bulb involve septal and ventral portions of the nose (A), whereas odorant representations in the lateral aspect of the bulb involve lateral portions of the nose (D). Because air exits ventrally from the rat nose, the prevailing flow of odorants is from dorsal to ventral across the olfactory epithelium. Odorants that are more soluble in mucosa (e.g., smaller molecules in a homologous series) therefore will tend to be absorbed more strongly to dorsal parts of the epithelium (magenta and green shading), while odorants that are less strongly absorbed (e.g., larger molecules in a homologous series) will be free to associate with more ventral epithelial regions (blue and red shading). The illustration of air flow in A is highly simplified; for example, in posterior and ventral regions of the nose, air actually flows from posterior to anterior, and under any given respiration condition, there are places in the nose where air velocity is zero (Kimbell et al., 1997). B, E: The structure of the bulb is shown as diagrams of coronal sections. The coloring of the glomeruli outlined in B shows the expected projection zones of similarly colored regions in A, whereas the coloring in E shows the expected projection of similarly colored regions in D. The topography of the epithelium-to-bulb projection is such that more dorsal epithelial regions are associated with more dorsal bulbar regions. C, F: The entire glomerular layer is represented as dorsal-centered charts. The medial acid-preferring response domain is outlined in C and colored to indicate the locations of glomeruli shaded in B. The corresponding lateral domain is outlined in F and shaded to show the locations of glomeruli in E. The relative absorption of odorants differing in carbon number together with the topography of the epithelium-to-bulb projection therefore could explain observed chemotopic progressions of glomerular responses from dorsal to ventral across the glomerular layer.

Responses to odorants are organized chemotopically in the rat olfactory bulb. Distinct odorant chemicals sharing certain molecular features (e.g., functional groups or aspects of hydrocarbon structure) or overall molecular properties (e.g., water solubility) stimulate overlapping, but distinct sets of glomeruli that are clustered within functionally defined domains distributed across the glomerular layer. These domains are illustrated as colored areas within a ventral-centered 2D plot of the glomerular layer in A, a dorsal-centered 2D plot in B, and various views of a 3D model of the surface of the glomerular layer in C. Chemotopically defined response domains are usually present as pairs, one on the medial aspect and one on the lateral aspect of the bulb. Arrows indicate the directions of systematic shifts in the location of responses that are observed for odorants of increasing carbon number along homologous series of straight-chained, saturated, aliphatic odorants.

Absolute measures of glomerular metabolic activity show increases in both intensity and area with increasing odorant concentration (top row), whereas responses relative to other responses in the glomerular layer are constant across odorant concentration (bottom row). Shown are contour charts averaged across three animals exposed to each of three concentrations of the odorant 1-pentanol (Johnson and Leon, 2000a). In the top row, 2DG uptake is expressed as a ratio of glomerular layer (GL) uptake to uptake occurring in a portion of the subependymal zone (SEZ), a region containing immature neurons that do not respond to odorants. Green and warmer colors indicate uptake that is greater than that detected in the same locations in animals exposed to air vehicle. In the bottom row, glomerular layer uptake at each location is expressed as a z score relative to the mean and standard deviation of all measurements across the glomerular layer. Green and warmer colors indicate uptake that is greater than the mean uptake across the glomerular layer in the same bulbs. Each color bin corresponds to one standard deviation.

Although individual odorants can stimulate much of the glomerular layer, relative levels of activity are very distinctive for different odorants. Color-coded contour charts are used to illustrate uptake evoked by seven different odorants. In the top row, responses are scaled as a ratio of uptake in each glomerular layer location to uptake measured in a consistent part of the subependymal zone, which is not expected to display odorant-dependent activity. Yellow or warmer colors indicate uptake that was above what was measured when rats were exposed to air vehicle only. In the bottom row, the same set of patterns are shown in a z score color scale, where yellow or warmer colors indicate uptake that is more than one standard deviation above the average uptake calculated across the glomerular layer. These patterns are very distinctive for different odorants. Outlined regions highlight consistent locations in the different charts. Whereas responses to different odorants in a given location do not seem as distinctive in the top row, the same locations are very differently active when viewed as relative (z score) responses in the bottom row.

Odorants can differ along many chemical dimensions. The odorant chemical in the center (pale yellow background) is 1-heptanol. Arrows indicate various incremental differences in structure, each of which produces a chemical of distinct steric and electronic properties. Although some dimensions, such as those involving changed or additional functional groups, result in greater differences in the molecular properties of the odorants than do other dimensions, almost all distinct chemicals evoke odors that can be discriminated by animals that are trained to do so. Studies of olfactory coding must account for both the breadth of odorant chemical structures that are detected by animals and the subtle discriminations that are possible between closely related compounds. Because of this dual challenge, the research must involve a great number of odorants.

Flexible odorants often activate clusters of glomeruli at high concentrations, whereas small or rigid odorants often activate isolated glomeruli in the same response regions. False-colored images of autoradiography sections indicate the relative distributions of [¹⁴C]2-deoxyglucose uptake within regions exhibiting primary responses (arrows) to each of the illustrated odorants, and enhanced-contrast images of adjacent cresyl violet-stained sections are overlaid to indicate the relationships between foci of uptake and individual glomeruli. A: Valeric acid and methyl valerate, which are flexible odorants capable of assuming a variety of conformations, evoke 2DG uptake over clustered sets of glomeruli. B: Smaller aliphatic compounds with the same functional groups, such as propionic acid and methyl acetate, cannot assume as many conformations and appear to activate isolated glomeruli in the same parts of the bulb. C: Rigid, cyclic structures such as cyclobutanecarboxylic acid and oxyoctaline formate also cannot assume multiple conformations and also appear to stimulate isolated glomeruli.

Center-surround lateral inhibition can narrow the molecular receptive range of mitral cells within chemotopically arranged glomerular response domains. Flexible aliphatic odorants often stimulate clusters of adjacent glomeruli (Figure 3). Although odorants differing by a single carbon in length activate overlapping sets of glomeruli, the strongest activation shifts systematically with carbon number in many domains (Figure 4). Because the most strongly activated mitral cells suppress the activity of their neighbors in a center-surround arrangement involving inhibitory periglomerular and granule cell interneurons, the systematic shifts in activity insure that mitral cells within a principal response domain are more selective for particular odorants within such a series than are the corresponding odorant receptors, sensory neurons, and glomeruli (Yokoi et al., 1985).

Weak, background responses to odorants can receive undue attention if the principal responses to the same odorants go unobserved. A: A color-coded contour chart of 2DG uptake (z scores) in response to the odorant octanal has been used to texture a 3-D model of the glomerular layer, the lateral aspect of which is illustrated here. The white arrowhead indicates a principal, ventrally located, response domain, where the uptake is more than 3 standard deviations above the mean uptake calculated across the entire glomerular layer. The black arrow indicates a relatively minor patch of uptake (green shading) that barely exceeds the average uptake across the glomerular layer. B: The 3-D model is rotated to emphasize the dorsal aspect of the bulb, which is the aspect that typically is studied in experiments using optical imaging techniques. C: Relative uptake within the region outlined in B is re-plotted as a surface chart with the same contour shading as in A and B. D: The surface plot in C is re-colorized to emphasize the largest responses visible in this region. This plot illustrates that weak “responses,” which may be due to odorant contaminants, can dominate the data collected using any method that either systematically samples from only a limited portion of the bulb or randomly samples only a few areas using only a few odorants. “Information” collected by such techniques can lead to incorrect conclusions involving broad response tuning and either broad or chaotic spatial distributions of odorant responses.

Very important odorant-evoked responses are spared following large bulbar lesions. A: Ventral-centered 2D charts show domains as in Figure 4 as well as color-coded z score responses to the odorants L-carvone and D-carvone averaged across six rats each (Linster et al., 2001). A semi-transparent, stippled overlay simulates a lesion removing 80% of the glomerular layer, including all of the lateral aspect. The area not covered by this overlay would be spared after such a lesion. The spared area includes at least a portion of five identified response domains as labeled in Figure 4. The spared area also includes the main differential response between L- and D-carvone (black arrow). B: This same spared region is shown on a 3D model of the glomerular layer viewed from a medial perspective. The representations of odorants that persist in this posterior part of the medial aspect of the bulb can explain the spared discriminatory powers of rats given such bulbar lesions.