This is a preprint of an article published in

The Journal of Comparative Neurology, 2000, 422:496-509.

Ó2000 Wiley-Liss

Modular representations of odorants in the glomerular layer of the rat olfactory bulb and the effects of stimulus concentration

Brett A. Johnson* and Michael Leon

Department of Neurobiology and Behavior, University of California,

Irvine, CA 92697-4550

Number of text pages: 45

Number of figures: 6 (1 color)

Number of tables: 2

Abbreviated title: Odorant coding in the olfactory bulb

Associate Editor: Joseph L. Price

Indexing terms: chemical senses, 2-deoxyglucose, 2-hexanone, mapping, odor, pentanal

*Correspondence to: Brett A. Johnson

Dept. of Neurobiology and Behavior

University of California

2205 BioSci II

Irvine, CA 92697-4550

Telephone: (949) 824-7303

Fax: (949) 824-2447

e-mail: bajohnso@uci.edu

Grant sponsor: NIDCD; Grant number: DC03545

Abstract

In order to study the mechanism whereby odorants are encoded in the nervous system, we studied the glomerular-layer activity patterns in the rat olfactory bulb evoked by closely related odorants from different chemical families. These odorants had a common straight-chain hydrocarbon structure, but differed systematically in their functional groups. Neural activity was mapped across the entire glomerular layer using the [¹⁴C]2‑deoxyglucose method. Group responses were averaged and compared using data matrices. The glomerular activity patterns that resulted from this analysis were comprised of modules. Unique combinations of modules were activated by each odorant, demonstrating what may be part of the neural code for odorants. Most of the modules were clustered together in the bulb, perhaps providing for enhanced contrast between related chemicals by means of lateral inhibition. We also determined whether changes in odorant concentration would affect spatial patterns of glomerular activity. Two odorants, pentanal and 2-hexanone, evoked different patterns at increased concentrations, with additional glomeruli being recruited at a great distance from glomeruli in which activity was evoked at lower concentrations. Humans report that both of these odorants change in perceived odor with increasing concentration. Three other odorants (pentanoic acid, methyl pentanoate, and pentanol) did not recruit new areas of glomerular activation with increasing concentration, and humans do not report a changed odor across concentrations of these odorants. The results suggest that changes in modular glomerular activity patterns could underlie altered odor perception across odorant concentrations, and they provide additional support for a combinatorial, spatially based code in the olfactory system.

At least one phase of odor coding appears to involve distinct spatial representations of odorants in the olfactory bulb. Odorants that evoke the perception of distinct odors also evoke distinct spatial activity patterns in the glomerular layer of the bulb (Stewart et al., 1979; Jourdan et al., 1980; Royet et al., 1987; Guthrie et al., 1993; Cinelli et al., 1995; Johnson et al., 1998, 1999). The spatial patterns evoked by an odorant are reliably seen across different individuals of the same species (Stewart et al., 1979; Jourdan et al., 1980; Friedrich and Korsching, 1997; Yang et al., 1998; Johnson et al., 1998, 1999; Galizia et al., 1999; Rubin and Katz, 1999). Modules of activity within these spatial patterns have been found to be correlated with important aspects of odorant chemistry (Friedrich and Korsching, 1997; Johnson et al., 1998, 1999; Rubin and Katz, 1999). For example, ethyl acetate, ethyl butyrate, isoamyl acetate, and isoamyl butyrate generate distinct patterns of glomerular activity in the rat olfactory bulb (Johnson et al., 1998). Within this group of aliphatic esters, those that shared aspects of chemical structure (e.g., an isoamyl group) stimulated common modules within the glomerular layer (Johnson et al., 1998), supporting the notion of a combinatorial coding of distinct molecular features of the odorants. Straight-chained aliphatic acids of progressively greater carbon number stimulated overlapping modules of the glomerular layer, and these modules shifted progressively across the layer in correlation with carbon number (Johnson et al., 1999). The progressive shifts in responses were interpreted as underlying the lateral inhibitory network that insures tuning of projection neurons to small differences in carboxylic acid carbon chain length (Mori et al., 1992; Imamura et al., 1992; Yokoi et al., 1995).

Our studies to date have involved mapping 2-deoxyglucose (2-DG) uptake across the entire glomerular layer after exposing animals to odorants possessing different hydrocarbon structures but the same functional groups (Johnson et al., 1998, 1999). Changes in hydrocarbon structure, especially in carbon chain length, allow facile comparisons between odorant structure and either neural activity or behavior (Döving, 1966; Mori et al., 1992; Imamura et al., 1992; Sato et al., 1994; Cometto-Muñiz et al., 1998; Laska and Teubner, 1998; Laska et al., 1999; Johnson et al., 1999; Malnic et al., 1999; Rubin and Katz, 1999). In the present study, we exposed animals to the five odorants shown in Table 1. These odorants share a four-carbon, straight-chained hydrocarbon structure, but they differ in their oxygen-containing functional groups. Such changes in functional groups alter the partial charge distributions of the odorant molecules, which might in turn affect their binding to particular odorant receptors in a manner distinct from alterations in hydrocarbon structure.

Whereas it seems clear that changes in odorant chemistry are correlated with activity in different glomerular modules, the effects of odorant concentration on these representations have not been well investigated. It would be particularly interesting if odorants that change in perceived odor across different concentrations were to evoke activity in different glomerular modules at different concentrations. Changes in odorant stimulus intensity can lead to changes in perceived odor quality in humans (Moncrieff, 1967; Theimer et al., 1977; Dravnieks, 1985; Gross-Isseroff and Lancet, 1988; Arctander, 1994; Pause et al., 1997). Although not all odorants display this phenomenon to the same extent, some of the changes can be very dramatic. In his compilation of the olfactory characteristics of numerous flavor and fragrance chemicals, Arctander reported concentration-dependent changes in the perceived odor of about 8% of the 2900 odorous compounds he surveyed (Arctander, 1994). Examples include numerous aliphatic aldehydes, which have fruity or floral odors at low concentrations but acrid and pungent or waxy odors at high concentrations (Arctander, 1994).

Changes in stimulus intensity also can change the perceived quality of the stimulus in other sensory systems (Gross-Isseroff and Lancet, 1988). Examples include a changed perceived pitch with increasing volume of auditory tones (Stevens, 1935; Snow, 1936), a similar change in perceived pitch with increasing amplitude of vibrating tactile stimuli (LaMotte and Mountcastle, 1975), and a perceived color shift with increasing intensity of particular wavelengths of light (Purdy, 1931).

If a spatial code underlies the perception of distinct odors, then concentration-dependent changes in the perceived odor of a chemical cue may be the result of a changed spatial pattern of glomerular response. Odors that do not change with concentration would be expected to have a largely stable pattern of glomerular activation. Indeed, spatial activity patterns across different odorant concentrations commonly are relatively constant (Stewart et al., 1979; Carmi and Leon, 1991; Guthrie and Gall, 1995; Cinelli et al., 1995; Joerges et al., 1997). Increased concentrations of odorants increase the amount of activity in responsive glomeruli (Stewart et al., 1979; Guthrie and Gall, 1995; Cinelli et al., 1995; Joerges et al., 1997; Friedrich and Korsching, 1997; Johnson et al., 1999; Rubin and Katz, 1999). Increased odorant concentrations also increase the area of focal glomerular responses (Cinelli et al., 1995; Friedrich and Korsching, 1997; Guthrie and Gall, 1995; Johnson et al., 1999; Rubin and Katz, 1999 Stewart et al., 1979). These local changes in the number of responsive glomeruli may be related to the spatial clustering of glomeruli of related specificity that presumably affords tuning of projection neurons by way of lateral inhibition (Yokoi et al., 1995; Johnson et al., 1999). While increased odorant concentration has been shown to recruit nearby glomeruli, there is no evidence that distant glomeruli are activated under these conditions, perhaps because these odorants have a constancy of odor quality perception with increasing concentration. Alternatively, such changes may exist for these odorants, but were not revealed by techniques that did not systematically map the activity in the entire bulb. Among the odorants we have chosen for this study, pentanal has been reported to change in odor with concentration, as have aliphatic ketones similar to 2-hexanone (Arctander, 1994).

MATERIALS AND Methods

Odorant dilutions and exposures

Odorant exposures were conducted following subcutaneous injections of [¹⁴C]2-DG, as described previously (Johnson et al., 1999). All odorants were purchased from Fisher Scientific (Tustin, CA). The listed purities of the odorants were as follows: pentanoic acid, 99%; methyl pentanoate, 99%; pentanol, 99%; pentanal, 98%; 2-hexanone, 98%. Odorants were volatilized using high-purity nitrogen to avoid oxidation, and the vapors then were diluted using a flow-dilution olfactometer and ultra zero-grade air. The level of dilution needed to produce a given vapor phase concentration was calculated from the vapor pressure of the odorant. Vapor pressures were estimated from boiling points and chemical classes using the equation of Hass and Newton (1975). Dilutions to achieve 25-ppm odorants were as follows: pentanoic acid, 1/8; methyl pentanoate, 1/537; pentanol, 1/106; pentanal, 1/1400; 2-hexanone, 1/628. The different dilutions of the different odorants are based on their inherent differences in volatility. A lower range of concentrations was chosen for pentanoic acid because 25 ppm was the highest concentration that could be used without depriving rats of oxygen and because previous studies had shown robust responses to concentrations as low as 7.2 ppm. Other chemical properties of odorants were estimated used Molecular Modeling Pro v.3 software (ChemSW Inc., Fairfield, CA).

The Irvine Animal Care and Use Committee (IACUC) approved all procedures. For most experiments, one Wistar rat from a given litter was exposed to air vehicle, and each of four additional rats from the same litter was exposed to a different concentration of the same odorant. For pentanoic acid, no air-exposed rat was included, but five concentrations of odorant were investigated. Consecutive exposures used increasing concentrations of odorant. Rats were tested between postnatal days 20 and 22. Rats from five litters were used for pentanoic acid, and from three litters for each other odorant.

2-DG procedure

Rats were given a subcutaneous injection of [¹⁴C]2-DG (Sigma, St. Louis, MO; 0.2 mCi/kg) and immediately were placed into a clean, 1-liter mason jar for 45 minutes. The odorant entered and the exhaust exited through the lid during that period. Following odor presentation, rats were immediately decapitated, their brains removed and then frozen in isopentane at ‑45°C. All tubing and connections leading to the exposure chamber were made of Teflon, Kynar, glass, or brass to minimize reactivity with the odorants. Clean exposure chambers and tubing were used for each odorant to minimize carryover from the previous odor.

Measurement of 2-DG uptake

Uptake of radiolabel was measured across the entire glomerular layer at systematic angle increments around 20-µm coronal bulb sections taken transverse to the long axis of the olfactory bulb (Johnson et al., 1999). Values from sections taken at equal intervals along the rostral-caudal axis were merged into arrays that then were either expanded or contracted to equalize the distance between anatomical landmarks encountered along that axis (Johnson et al., 1999).

As in our previous study (Johnson et al., 1999), the landmarks included the first section possessing an external plexiform layer, the first section possessing an accessory olfactory bulb, and the last section possessing a mitral cell layer on the medial aspect. For a bulb of modal size, adjacent values in these arrays were separated by 120 µm. Final standardized arrays contained 44 sections, with the largest sections yielding 80 measurements along the glomerular layer. The first accessory olfactory bulb-containing section was always adjusted to be section 25. Measurement numbers in each section were adjusted such that the ventral-most measurement was at the same position in all sections (measurement number 39).

Gray scale values of film density were converted to nCi/g of ¹⁴C by comparison with standards. The arrays from the left and right bulb of a given rat were averaged. These average arrays then were subjected to different transformations depending on the analysis to be performed.

In order to compare entire maps across conditions, we calculated a pattern dissimilarity index, as described previously (Johnson et al., 1999). This calculation subtracts one z score-standardized array from another, followed by squaring all values in the difference array, averaging the squares across the difference array, and then taking the square root of this mean square. Thus, each pair of arrays contributes a single value that is greater when the patterns are more dissimilar.

Centroid calculations

To compare statistically the spatial distributions of glomerular activity produced by different odorant concentrations, it was useful to describe the pattern in each animal using a single value, which avoids the problem of multiple statistical comparisons. We chose to compare centroids of activity, which describe in two coordinates the location within an array of values where the sums of values on all sides are equal. Statistical comparisons of centroids have been used previously in analyzing spatial distributions of activity across the olfactory epithelium (Youngentob et al., 1995) and bulb (Johnson et al., 1999).

Arrays from air-exposed controls within the same litters were subtracted from each other array to correct for any litter-specific contaminants (e.g., those carried over from home cages). In the case of pentanoic acid, the average array from rats exposed to the lowest odorant concentration was subtracted because this condition was found not to be distinguishable from air. Then, group means (across different rats exposed to the same concentration of the same odorant) and grand means (across all rats exposed to any concentration of the same odorant) of each coordinate of the centroid were calculated. For each coordinate and each mean, we calculated the difference between the value in an individual rat and the value of the mean. The differences in the two coordinates were squared and summed. The square root of this sum represented the distance between the centroid in the individual rat and the mean centroid. These distances then were used to calculate F and t values for the appropriate statistical tests.

Results

Patterns involve glomerular modules.

Figure 1 shows the average patterns of uptake for each odorant exposure after subtraction of appropriate air control patterns. The patterns obtained for different concentrations of pentanoic (valeric) acid in the present study are very similar to the patterns we obtained using two different sets of animals in our previous study (Johnson et al., 1999), which is evidence for the reliability of our mapping procedure. The average maps that are shown in Figure 1 also were quite representative of the maps seen for each bulb of each individual animal contributing to the averages.

By comparing the patterns evoked by the different odorants, it became clear that particular sets of odorants evoked overlapping activity. Often, an area of activity evoked by a given odorant overlapped only partially with an area of response evoked by another odorant. For example, the small patch of high activity in the lateral bulb that was evoked by 25-ppm 2-hexanone (Fig.1) overlapped only with the more dorsal part of the caudal and lateral activity evoked by either 12.5-ppm pentanoic acid or 25-ppm pentanal (Fig.1). This partial overlap implied that the region of response to pentanoic acid or 25-ppm pentanal might be comprised of two smaller modules conveying distinct information about the odorant stimulus. Indeed, the caudal and lateral region of response to pentanoic acid and 25-ppm pentanal showed direct signs of being comprised of two smaller components of activity. Each time such a partial overlap was detected in the lateral sector, a similar observation could be made in the medial sector, consistent with the medial/lateral pairs of modules that displayed similar specificity in our previous studies (Johnson et al., 1998, 1999).

Often, the paired modules that distinguished one odorant from another odorant also distinguished one concentration of a given odorant from another concentration. For example, the extreme rostral modules of activity in the lateral and medial bulbs of rats exposed to pentanoic acid overlapped exactly with the rostral modules evoked by higher, but not lower, concentrations of pentanal (Fig. 1).

After a careful alignment of each odorant pattern with each other odorant pattern, a total of eight lateral modules were identified. These lateral modules are illustrated and labeled with lower case letters in Figure 2. The medial modules corresponding to each of these lateral modules are labeled with corresponding upper case letters (Fig. 2). One module of response in the medial bulb, module I, appeared not to have a lateral equivalent.

To demonstrate how amounts of activity within these modules differed across odorants and across concentrations, we standardized arrays within each experiment by calculating the ratio of glomerular layer uptake/subependymal zone uptake (Coopersmith and Leon, 1984; Johnson and Leon, 1996) and by subtracting appropriate air-exposed control arrays. This standardization is distinct from that used for Figure 1, which involves a z score transformation relative to the mean and standard deviation of all values measured throughout the glomerular layer. Z-score transformations also correct for different amounts of 2-DG injected, but high activity in one part of the glomerular layer can lower the values calculated at any other part of the bulb.

Average values of glomerular layer uptake/subependymal zone uptake were calculated across the 17 modules for each animal in the present study. Mean values for each odorant condition are illustrated by the diameters of black circles in Figure 3. These values have been further standardized for each odorant relative to the largest value obtained at any concentration of that odorant. Therefore, the diameters of the circles can be compared directly within a given odorant, but only for differences in pattern across odorants. The outer and inner diameters of each gray annulus denote the mean plus and minus the standard error, respectively. The consistency of the results from animal to animal is evidenced by the fact that the thickness of the black line usually obscures the gray annulus depicting the variance across animals (Fig. 3).

A number of important conclusions concerning odorant representations are evident from Figure 3. One of the most striking observations is the similarity of the specificity of corresponding lateral and medial modules. With very few exceptions, lateral and medial modules were evoked to a similar degree at any given concentration of an odorant, which gives the appearance of similar patterns of circle size on the left and right sides of Figure 3.

Differences between odorants are perhaps more apparent in the modular representation of 2-DG uptake (Fig. 3) than they were in the complete maps shown in Figure 1. Each odorant appears to activate a unique combination of glomerular modules. For example, the relatively greater activity in modules a/A, d/D, and e/E in the representation of pentanoic acid distinguishes it from other odorants. The nearly equal utilization of posterior modules e/E, f/F, g/G, and h/H distinguishes methyl pentanoate from most odorants. The participation of most modules, but especially f/F, a/A, and b/B makes pentanol unique. Qualitatively, some odorants appeared to overlap more in their use of modules than did others. For example, the relatively high activation of posterior modules (e-h/E-H) at 25-ppm 2-hexanone is similar to the relative activation of the same modules in the representation of methyl pentanoate.

Activity patterns change with odorant concentration

Evoked patterns of activity were readily apparent for each of the odorants at the higher concentrations studied (Fig. 1). For three of the five odorants, the pattern evoked at high odorant concentrations was largely similar to the pattern encountered at the lowest concentration that evoked a detectable pattern. In contrast, the patterns evoked by pentanal and 2-hexanone appeared quite distinct at low and high concentrations [compare the maps at 25 parts per million (ppm) to the maps at 250 ppm]. For these two odorants, the modules of response dominating the pattern at 25 ppm became minor contributors at 250 ppm (black arrows in Fig. 1). With increasing concentration of these two odorants, multiple new modules of response were evoked (white arrows in Fig. 1) and some of these new modules were located quite distant from the original responsive modules.

As shown in Figure 3, most modules showed increased 2-DG uptake with increasing odorant concentration, as has been reported previously (Stewart et al., 1979; Carmi and Leon, 1991; Johnson et al., 1999). However, there were a few interesting exceptions. Animals exposed to 25-ppm pentanoic acid actually displayed less 2-DG uptake in all modules relative to animals exposed to 12.5-ppm pentanoic acid. With an increase in the concentration of 2-hexanone from 75 ppm to 250 ppm, there were decreases in the amounts of uptake in modules d/D, e/E, g/G, and h/H. With the same step in concentration of 2-hexanone, uptake in modules b/B, c/C, and I clearly increased. The pattern change at increasing concentrations of pentanal is evident both as an increased activation of modules e/E relative to modules d/D and as a new activation of modules a/A, b/B, f/F, and h/H. The pattern change for 2-hexanone can be seen as a shift from a distributed pattern of posterior modules e-h/E-H to a pattern emphasizing modules c/C, e/E, and f/F.

Before determining the statistical significance of apparent changes in pattern with odorant concentration, it was necessary to determine which of the odorant concentrations yielded patterns distinguishable from air. We first calculated an average air pattern for the current study and then determined indices of pattern dissimilarity between this pattern and the average air patterns obtained in seven prior experiments in this laboratory. The distribution of these seven values is shown in Figure 4A. The mean dissimilarity index was 0.660, and the standard deviation was 0.052.

We then calculated indices of pattern dissimilarity between each of the 21 odorant conditions in the present study and the average air pattern for this study. The distribution of these values is shown in Figure 4B. The latter distribution was bimodal, with the lower mode falling within one standard deviation of the mean value obtained when the average air pattern was compared to other average air patterns (Fig. 4A). Thus, the patterns obtained in the odorant conditions contributing to this lower mode are consistent with air patterns, and it was decided that these patterns should not be included in statistical analyses of changes in odorant-evoked patterns at different concentrations. The odorant conditions that were eliminated all involved low odorant concentrations: 1.6-ppm pentanoic acid (not shown in Fig. 1), 7.5- and 25-ppm methyl pentanoate, 2.5-ppm pentanol, 7.5-ppm pentanal, and 7.5-ppm 2-hexanone. The odorant conditions judged to give patterns different from air are marked with asterisks in Figure 1.

To determine whether concentration-dependent changes in pattern were statistically significant for any of the odorants, we divided arrays at the ventral extreme to produce two sectors, one lateral and one medial. These sectors were considered separately because our previous experiments have shown that the medial and lateral aspects of the bulb contain separate, matching representations of aliphatic ester and aliphatic acid odorants (Johnson et al., 1998, 1999). We then calculated centroids of uptake in each of these sectors. The centroids were determined as a rostral-caudal coordinate and a dorsal-ventral coordinate. Although centroids are a crude measure of the overall patterns evoked by odorants, shifts in centroids are a good indication of changes in pattern that involve an altered distribution of values in one direction or another across the arrays (Youngentob et al., 1995; Johnson et al., 1999).

Centroids were calculated for each animal exposed to an odorant at a concentration that was determined to evoke a pattern distinct from air. In Figure 5, the centers of ellipses indicate mean positions of the centroids, and the lengths of the axes of each ellipse indicate the standard errors across different animals in the two dimensions used to calculate the centroids. Ellipses located more to the left in Figure 5 indicate more rostral centroids and those located towards the top indicate more dorsal positions in the lateral bulb or more ventral positions in the medial bulb. Statistical analysis revealed that the centroids were different across concentration in the medial sector of the bulb for pentanal and in both the lateral and the medial sectors for 2-hexanone (Table 2). As shown in Figure 5, bottom, the centroids calculated in the medial sector for 2-hexanone (black ellipses) shifted to more dorsal positions with increasing odorant concentration. A similar observation can be made for centroids calculated for pentanal (gray ellipses) in the medial sector (Fig. 5, bottom). In the lateral sector, centroids calculated for 2-hexanone (black ellipses) also shifted to more dorsal positions with increasing odorant concentration (Fig. 5, top). Although not statistically significant, centroids calculated for pentanal also showed a tendency towards more dorsal positions with increasing concentration (gray ellipses, Fig. 5, top). No other odorant was found to give statistically significant changes in centroids across concentration (Table 2).

Pattern differences across concentrations can be as large as differences across odorants

To obtain a quantitative measure of the relative similarities in pattern between different odorants and different concentrations, we subjected each set of the eight mean values obtained for the lateral modules to pattern dissimilarity analysis (Fig. 6). This analysis involved pair-wise comparisons of each average odorant condition to each other average odorant condition after a z score-transformation of the values for the eight modules. We restricted this analysis to those odorant concentrations judged to evoke patterns different from those seen in air-exposed animals. To illustrate the distribution of the large number of resulting comparisons, we have assigned to each pattern dissimilarity index a gray-scale value that is proportional to the magnitude of the difference in the activity patterns. The pattern of dissimilarities takes on the appearance of a patchwork, with darker squares indicating large dissimilarities and lighter squares indicating large similarities (Fig. 6).

As shown by black arrows in Figure 6A, the modular patterns evoked by two different concentrations of the same odorant sometimes were more different than those evoked by different odorants. For example, 25-ppm and 250-ppm 2-hexanone resulted in higher values of pattern dissimilarity than did comparisons between 25-ppm 2-hexanone and either 75-ppm or 250-ppm methyl pentanoate. That is, low concentrations of 2-hexanone were found to evoke a pattern of activity more similar to that evoked by methyl pentanoate than that evoked by higher concentrations of 2-hexanone. Similarly, comparisons of 25-ppm and 250-ppm pentanal resulted in higher indices of pattern dissimilarity than did comparisons between 250-ppm pentanal and either 75-ppm or 250-ppm methyl pentanoate (white arrows in Fig. 6A).

We also generated patchwork diagrams representing pattern dissimilarity indices that were calculated across the eight medial modules corresponding to the eight lateral modules (Fig. 6B). The analysis of the medial modules produced the same conclusions about similarities between methyl pentanoate and either low concentrations of 2-hexanone or high concentrations of pentanal (arrows in Fig. 6B). Indeed, many other details of the patchwork representing lateral modules were replicated in the analysis of the medial bulb. For example, the greatest dissimilarities were found in comparisons of pentanoic acid and 2-hexanone in both lateral and medial modules (see the upper right corners of Fig. 6A and 6B). Together, these data provide even greater evidence for the presence of two nearly identical representations of odorants in the bulb, one in the lateral, and one in the medial aspect.

The patchwork generated from pattern dissimilarity calculations performed across the approximately 2500 values in each average array is shown in Figure 6C. All of the same conclusions we obtained for the sets of eight modules also were obtained from comparisons of entire arrays. Thus, for the odorants and concentrations studied here, most of the information regarding different amounts and locations of 2-DG uptake across the entire glomerular layer of the olfactory bulb could be condensed to eight values representing eight glomerular modules. The relative homogeneity in pattern dissimilarity values across Figure 6C (in comparison to either Figure 6A or 6B) suggests that activity outside of the modules did not differ for the odorants in this study. The numerous low-difference values calculated at positions outside of the modules lowered the average dissimilarity calculated across the arrays.

Discussion

A modular basis of odorant representations

The differences in spatial patterns evoked by different odorants and different odorant concentrations in the present study could be described in terms of differential activity in discrete glomerular modules. These modules were defined operationally by comparing the simplest average odorant-evoked patterns with increasingly complex patterns. Often, modules that distinguished one odorant from another odorant also distinguished one concentration of a given odorant from another concentration of the same odorant. When the odorant-evoked responses of individual rats were analyzed with respect to these modules, the variance across animals exposed to the same odorant condition was low, suggesting that the modules adequately described patterns present in individual olfactory bulbs. Mathematical comparisons of pattern differences using only information from the modules recapitulated pattern differences calculated across the entire glomerular layer, suggesting that the modules may contain sufficient information to distinguish between the odorant conditions compared in the present study. Together, these data indicate that modular glomerular activity of odorants may represent part of an olfactory code.

Despite the interpretative power gained through the analysis of glomerular modules, it is difficult to derive from our data the precise anatomical composition of these modules. The sizes of the modules in our average maps depend on multiple factors. These factors include the number of glomeruli activated in a given bulb, the size of the activated glomeruli, and the uncertainty in the location of a glomerulus from one bulb to another. The absence of a sharp border of 2-DG foci, as well as the possibility that 2-DG uptake might extend beyond the borders of activated glomeruli, also limit our ability to define the number of glomeruli associated with a given module. In individual bulbs from rats exposed to pentanoic acid or 25-ppm pentanal, the uptake within modules d/D and e/E appeared to be associated with as few as one or two glomeruli (data not shown). As such, these modules are reminiscent of certain modules or “fields” activated by isoamyl butyrate and acetic acid in previous studies (Johnson et al., 1989, 1999). Activation of a single glomerulus (or of a few adjacent glomeruli) may be explained by the activation of a single olfactory receptor protein, because sensory neurons expressing the same olfactory receptor gene converge in their projections to one or a few adjacent glomeruli in the olfactory bulb (Mombaerts et al., 1996; Ressler et al., 1994; Vassar et al., 1994; Wang et al., 1998).

Previous analyses indicated that pentanoic acid can activate as many as 35 glomeruli in the regions of the bulb defined as modules a and A in the present study (Johnson et al., 1999). Modules c and C that were activated by 2-hexanone in the present study also appeared to include a large number of glomeruli. Because at least some modules reflect the activity of multiple glomeruli, it is possible that other odorants could activate only portions of each of the currently defined modules. Therefore, it is possible that the modules discussed in the present study are themselves composed of smaller modules. Indeed, a further investigation of the patterns evoked by aliphatic acids in our previous study (Johnson et al., 1999) indicated that propionic acid and caproic acid stimulate the dorsal and ventral aspects, respectively, of modules a and A, and that their activity within these modules only barely overlaps.

All of the glomerular modules defined in the present study were activated by more than one odorant (Fig. 3). The most extreme examples were modules e, E, and I, which were activated by all five of these odorants, although they were not part of the patterns observed for acetic or propionic acid in our previous study (Johnson et al., 1999). Thus, odorants can be distinguished by the combination of modular units that they activate. The combinatorial use of glomerular modules thus supports the multiple receptor site, multiple molecular feature model originally proposed by Polak (1973). Subsequently, this attractive notion was suggested by other groups (Lancet, 1986; Kauer and Cinelli, 1993; Mori and Yoshihara, 1995; Johnson, et al, 1998; Malnic, et al, 1999). Combinatorial odor coding has been supported by psychophysical data (Theimer et al., 1977; Gross-Isseroff and Lancet, 1988; O’Connell et al., 1994), the finding of overlapping specificities of olfactory sensory neurons (Malnic et al., 1999), and evidence for modular representations of odorants in the glomerular layer (Cinelli et al., 1995; Friedrich and Korsching, 1997; Johnson et al., 1998, 1999). Our present data support the notion of combinatorial coding of odorants wherein distinct functional groups of molecules contribute to the stimulation of overlapping but distinct sets of glomerular modules.

The extensive overlap in the use of glomerular modules across odorants differing in functional groups resembles the overlap in responses of individual sensory neurons (Sato et al., 1994; Malnic et al., 1999) and individual bulbar mitral/tufted cells (Mori et al., 1992; Imamura et al., 1992). Structural modeling of odorant receptor proteins suggests that hydrogen bonding between polar receptor side chains and oxygen-containing functional groups of odorant ligands may be important for the binding of certain odorants (Singer and Shepherd, 1994; Singer et al., 1995, 1996). Hydrophobic interactions between apolar receptor side chains and odorant hydrocarbons may be more important in determining the specificity of these odorant-receptor interactions (Singer and Shepherd, 1994; Singer et al., 1995, 1996; Pilpel and Lancet, 1999).

A majority of the glomerular layer (67%) was not activated by any of the odorants in the present study. Because these glomeruli are probably activated by unique molecular features possessed by other odorants, there is a great opportunity for other combinations of modules that would allow the encoding of a vast number of odorants. Indeed, odorants in our previous studies evoked patterns that both overlapped with the modules in the present study and indicated the presence of other modules (Johnson et al., 1998, 1999).

Various aliphatic esters stimulated broad expanses of the posterior olfactory bulb in our previous study, and centroids of response in this region were found to shift rostrally with increasing odorant hydrophobicity (Johnson et al., 1998). A similar finding was obtained for a series of aliphatic acids (Johnson et al., 1999). Numerous other odorants also had been reported to stimulate this region of the bulb. All of these observations prompted us to hypothesize that this part of the bulb may receive projections from lower specificity receptors monitoring the overall hydrophobicity of odorants. Given our present data, this hypothesis appears no longer tenable. Individual odorant molecules in the present study activated only a few glomeruli within this region, indicating a high specificity of these glomeruli for particular odorants. Also, the relative locations of the activated glomeruli had no relationship to the hydrophobicity of the odorants (the predicted relative hydrophobicities of the compounds are methyl pentanoate >> pentanal > pentanoic acid » pentanol > 2-hexanone).

Clustering of modules responding differentially to odorants with distinct functional groups

Many of the modules that responded to odorants with different functional groups are located adjacent to one another in the midlateral and midmedial aspects of the glomerular layer (Fig. 2). Because mitral cell dendrites extend about 0.5 mm in the rat (Mori, 1987), these modules are probably within each other’s range of lateral inhibition. Thus, if pentanoic acid stimulates module e, but not adjacent module f, and if pentanol stimulates both module e and module f, then projection neurons associated with module f could be activated by pentanol but strongly inhibited by pentanoic acid. Thus, lateral inhibition may serve to enhance the differences in activity of certain projection neurons during exposures to odorants bearing similar functional groups.

Paired lateral and medial representations of odorants

For every module of response identified in the lateral aspect of the glomerular layer, there was a module of very similar specificity identified in the medial aspect. The similarity was evident in terms of both the relative activation across different odorants and the degree of activation at different concentrations of a given odorant. This observation greatly extends our results from previous studies of aliphatic ester and acid odorants (Johnson et al., 1998, 1999). As in the previous studies, the medial module of the pair was located more caudally and ventrally than the lateral module. The relative locations of the modules are the same as the relative locations of lateral and medial glomeruli receiving projections from sensory neurons expressing the same olfactory receptor gene (Mombaerts et al., 1996; Ressler et al., 1994; Vassar et al., 1994; Wang et al., 1998). Possible reasons for two representations of odorant molecular features in a single bulb have been discussed previously (Johnson et al., 1999).

One module of response in the medial aspect of the bulb (module I) did not appear to have a lateral equivalent. It was stimulated to some extent by all of the odorants in the present study, and further analysis of patterns in previous studies indicated that it probably was stimulated by isoamyl acetate and isoamyl butyrate, as well (Johnson et al., 1998). Module I also was remarkable for its extreme ventral and posterior position; it was associated with the most ventral glomeruli of the sections in which it was detected (not shown).

Relative advantages of different imaging techniques

With the advent of new imaging techniques, limitations of the 2-DG technique recently have been discussed (Yang et al., 1998; Rubin and Katz, 1999). These limitations include the relatively long odorant exposure periods employed and the resulting loss of temporal information. However, techniques capable of temporal measurements have indicated that spatial patterns of glomerular activity are constant across time in the rat (Guthrie and Gall, 1995; Rubin and Katz, 1999), suggesting that the absence of temporal data in a 2-DG study may not be an important limitation. It also is the case that intermittent sniffing of an odorant for a total of only 50-150 seconds distributed over a 45-minute exposure time results in the same pattern of 2-DG uptake as does a continuous 45-minute odorant exposure (Slotnick et al., 1989).

The animals being studied using the 2-DG technique are neither restrained nor anesthetized, and therefore are moving and breathing naturally. The anesthesia used in most of the other imaging techniques reduces natural sniffing and almost certainly depresses neural activity. There also is no prior surgery in the 2-DG method. Therefore, 2-DG uptake may represent neural activity occurring in a more natural context than that present during the use of the other techniques.

Even more importantly, mapping of 2-DG uptake probably provides the only available method to access the entire olfactory bulb at a resolution allowing the standardized measurement of every active glomerulus. While the spatial resolution of the 2-DG method is sufficient to identify a single activated glomerulus in a section (Johnson et al., 1998, 1999), this resolution is not maintained when maps from multiple bulbs are averaged. Nevertheless, the anatomical standardization of our arrays across different animals allows statistical comparisons that ensure the reliability of the results from animal to animal. Statistical comparisons across animals have not yet been applied in studies using other imaging techniques. Once such across-animal comparisons are made, the averaging required also will result in a loss of apparent spatial resolution.

In optical imaging, information can come only from parts of the structure accessible to the camera. In rats, this means that data collection is limited to the dorsal surface of the bulb, representing only about 10% of all glomeruli (Rubin and Katz, 1999). In the absence of other data, such as that obtainable from the 2-DG technique, one would not know whether any responses detected on the dorsal surface represent major components of the overall spatial pattern of activity. Indeed, the dorsal responses to pentanal that were measured recently using optical imaging (Rubin and Katz, 1999) probably corresponded to module a in our study. This module was a relatively minor contributor to our pentanal-evoked patterns and was not seen at the lowest concentration of the odorant (Fig.1). The majority of the response to pentanal in our study occurred in the midlateral and midmedial portions of the bulb, which would not have been accessible to optical imaging.

Another technique that has been applied recently to olfactory coding is functional magnetic resonance imaging (fMRI; Yang et al., 1998). The fMRI technique shares with 2‑DG mapping the ability to detect activity throughout the dorsal-ventral extent of the bulb, and, when improved in its spatial resolution and standardized across different animals, it should yield complementary results (Yang et al., 1998). Both fMRI and optical imaging have the advantage of being able to expose individuals to many odorants or to repeated presentations of the same odorant, and to monitor these responses across time. Optical imaging has the further advantage of combining glomerular-resolution activity with single-unit recordings and labeling of recorded neurons (Rubin and Katz, 1999). It thereby offers the promise of understanding the relationships between glomerular activity and primary responses of mitral and tufted cells to specific odorants.

Finally, odorant exposures lead to increased transcription of immediate early gene mRNAs and translation of their protein products, such as c-Fos and Egr-1 (Onoda, 1992; Guthrie et al., 1993; Sallaz and Jourdan, 1993; Johnson et al., 1995). Different odorants can be shown to generate distinct qualitative patterns using in situ hybridization for c-fos mRNA (Guthrie et al., 1993). In situ hybridization or immunohistochemistry directed against immediate early gene products allow cellular resolution of the increased activity. Nevertheless, amounts of activity can not be compared readily across different animals. Each of these procedures involves incubations of sections in numerous solutions, which in practice can result in variable backgrounds and staining intensities across different sections and even across different parts of the same section. There are no good methods for standardizing these signals across different incubations, because the background and the signal can vary independently. Combining sections from numerous animals in single incubations and setting strict definitions for counting individual stained cells can be effective for certain purposes (Johnson et al., 1995), but the number of individual animals and sections needed for a study such as the present one makes this approach extremely impractical. Therefore, the analysis of immediate early gene products is not well suited for mapping responses to closely related odorants.

Relationships between odorant concentration and levels of glomerular activity

For most odorants in the present study, increased concentration was correlated with increased 2-DG uptake in all glomerular modules (Fig. 3). Increased glomerular activity with increasing odorant concentration is a common observation across many studies, regardless of the technique used to measure activity (Stewart et al., 1979; Guthrie and Gall, 1995; Cinelli et al., 1995; Joerges et al., 1997; Friedrich and Korsching, 1997; Johnson et al., 1999; Rubin and Katz, 1999). The increased glomerular activity probably results from increased activity in olfactory sensory neurons that have projections converging into those glomeruli.

Two apparent exceptions to the rule of increased 2-DG uptake with increasing odorant concentration involved changes from 12.5-ppm to 25-ppm pentanoic acid and from 75-ppm to 250-ppm 2-hexanone. At 25-ppm pentanoic acid, all modules were decreased in their 2-DG uptake when compared to 12.5 ppm. A likely explanation for this uniform decrease is a decreased inspiration of pentanoic acid with this concentration step. Pentanoic acid activates trigeminal nerve endings in the respiratory epithelium of the nose (Cometto-Muñiz et al., 1998). Trigeminal activation causes reduced respiration to minimize exposure to chemical irritants, and the decrease in respiration is dependent on the concentration of the irritant (Alarie, 1966, 1973; Chang et al., 1981; Barrow and Steinhagen, 1982).

A more complex explanation must apply to the decreases in 2-DG uptake that occur in only certain modules between 75-ppm and 250-ppm 2-hexanone. This concentration step was associated with increased 2-DG uptake in modules b/B, c/C, and I, but decreased uptake in modules d-h/D-H (Fig. 3). Possible explanations range from a change in sniffing that redirects air flow to particular portions of the olfactory epithelium (Youngentob et al., 1987), to the presence of negative cooperativity in the binding of certain odorant ligands to receptors responsible for activity in modules d-h/D-H. We have no reason to prefer the latter hypothesis at this time, but if true, it may have profound implications for odor coding.

Changed spatial patterns of glomerular activity with odorant concentration

We have demonstrated that different concentrations of pentanal and 2-hexanone can produce different spatial patterns of activity in the glomerular layer of the rat olfactory bulb. In both cases, higher concentrations evoked activity in parts of the bulb that were not activated at lower concentrations.

Activation of additional olfactory sensory neurons at higher odorant concentrations has been attributed to the recruitment of distinct classes of odorant receptors that have lower affinity for the odorants (Malnic et al., 1999), and this recruitment has been proposed to underlie changed odor perception (Malnic et al., 1999). Our observations of changed spatial patterns of glomerular activity are consistent with this proposal. The activation of new glomeruli at higher odorant concentrations likely indicates the recruitment of new olfactory receptor proteins, since sensory neurons expressing the same receptor gene project to as few as one lateral and one medial glomerulus (Ressler et al., 1994; Vassar et al., 1994; Mombaerts et al., 1996; Wang et al., 1998).

The recruitment of olfactory sensory neurons at higher odor concentrations, however, may not always result in a different perceived odor; typically, the odor just increases in intensity. The consequences of activating different olfactory sensory neurons may depend on the spatial arrangements of the glomeruli to which the sensory neurons project. Glomeruli responding maximally to aliphatic acids of a given carbon chain length appear to be located near glomeruli that respond maximally to acids of slightly different carbon chain lengths (Johnson et al., 1999), suggesting that sensory neurons expressing receptors of similar selectivity may project to neighboring glomeruli (Singer et al. 1998; Tsuboi et al., 1999). Similar binding specificities of neighboring glomeruli also would explain why higher concentrations of odorants increase the area of focal glomerular responses detected using 2-DG uptake (Stewart et al., 1979; Johnson et al., 1999), in situ hybridization for c-fos mRNA (Guthrie and Gall, 1995), voltage-dependent dye recording (Cinelli et al., 1995; Friedrich and Korsching, 1997), or optical recording (Rubin and Katz, 1999). The increased area of response may reflect the recruitment of adjacent glomeruli that are maximally responsive to odorants of slightly distinct chemistries, but that can respond to higher concentrations of suboptimal stimuli (Johnson et al., 1999).

Spatial clustering of glomeruli with similar specificities has been proposed to invoke a lateral inhibitory network that tunes projection neurons to specific odorant carbon chain lengths (Yokoi et al., 1995; Friedrich and Korsching, 1997; Johnson et al., 1999). Neighboring mitral and tufted cells can inhibit one another via inhibitory granule cells, and this type of inhibition may be greater than the lateral inhibition at the glomerular level. Therefore the increased area of glomerular activation may not be reflected at the level of projection neurons, and the perception of odor quality would not change as odor intensity increases.

In the present study, higher concentrations of pentanal and 2-hexanone stimulated glomeruli distant from those activated at lower concentrations. For example, the greatest activity in module A, which was evoked by high concentrations of pentanal, was located about 2.6 mm from module E, the nearest module activated by 25-ppm pentanal. The greatest activity in module C, which was activated by high concentrations of 2-hexanone, was located about 2.0 mm from module E, the area showing the greatest activity at 25-ppm 2-hexanone. Lateral inhibition initiated by the activation of a single glomerulus should be confined to the area occupied by the basal dendrites of the mitral cells associated with the activated glomerulus. These dendrites extend about 0.5 mm in the rat (Mori, 1987). Therefore, certain glomeruli evoked at high, but not low, concentrations of pentanal and 2-hexanone are likely to be outside the range of lateral inhibition from the originally activated glomeruli. The consequent activation of distinct sets of mitral cells in different parts of the olfactory bulb may contribute more to a changed odor perception than would the recruitment of nearby glomeruli.

Possible relationships between spatial patterns of activity and odor perception

Although there is as yet no data on odor perception in rats across different concentrations of the odorants included in the present study, human observations are consistent with a relationship between the different spatial patterns of activity found here and the perception of different odors. Humans report that low concentrations of pentanal evoke a dry-fruity or nutty odor, whereas high concentrations evoke a more acrid and pungent odor (Arctander, 1994). Humans also report that aliphatic ketones similar to 2-hexanone smell ethereal, spicy, or fruity at low concentrations, but have more pungent, “chemical”, or solvent-like odors at high concentrations (Arctander, 1994). Although the perception of pungency at high odorant concentrations may be due in part to trigeminal nerve activity (Alarie, 1966), our data suggest the possibility that glomerular activity patterns also may participate in producing different odors.

Our data further indicate that certain concentrations of either pentanal or 2-hexanone can generate spatial patterns of activity more similar to those evoked by methyl pentanoate than to those evoked by other concentrations of pentanal or 2-hexanone, respectively. The spatial coding hypothesis would predict that rats might perceive a similarity between the odors of these different chemicals if tested at the appropriate concentrations. This prediction has not yet been tested in rats, but methyl pentanoate and other aliphatic esters have a fruit-like odor to humans (Dravnieks, 1985; Arctander, 1994), and this same descriptor is applied to certain concentrations of aldehydes and ketones. These data raise the possibility of a similarity in odor perception between humans and rats. Indeed, similarities in odor perception across species have been suggested recently for honeybees, monkeys, and humans (Laska and Teubner, 1998; Laska et al., 1999).

Experiments currently underway to test the specific predictions about odor perceptions that arise from our pattern comparisons in rats should go far towards evaluating the spatial coding hypothesis of odor perception. Thus far, our data support the notion that different molecular features of odorants first are isolated by activating members of a large family of olfactory receptors. These signals then are amplified by convergence in specific glomeruli, sharpened by lateral inhibition in the bulb, and then presumably reconstructed as an odor perception in the olfactory cortex (Johnson et al., 1999).

Acknowledgments

We thank Edna E. Hingco, Zhe Xu, Zsuzsanna B. Najbauerné, and Keith L. Pham for excellent technical assistance, Dr. Joseph Najbauer for helpful comments, and Dr. Cynthia Woo for thoughtful discussions and a critical reading of the manuscript.

Literature Cited

Alarie Y. 1966. Irritating properties of airborne materials to the upper respiratory tract. Arch Environ Health 13:433-449.

Alarie Y. 1973. Sensory irritation by airborne chemicals. CRC Crit Rev Toxicol 2:299-363.

Arctander S. 1994. Perfume and Flavor Chemicals (Aroma Chemicals). Carol Stream, IL: Allured Publishing Company.

Barrow CS, Steinhagen WH. 1982. Sensory irritation tolerance development to chlorine in F-344 rats following repeated inhalation. Toxicol Appl Pharmacol 65:383-389.

Carmi O, Leon M. 1991. Neurobehavioral responses of neonatal rats to previously experienced odors of different concentrations. Dev Brain Res 64:43-46.

Chang JCF, Steinhagen WH, Barrow CS. 1981. Effect of single or repeated formaldehyde exposure on minute volume of B6C3F1 mice and F-344 rats. Toxicol Appl Pharmacol 61:451-459.

Cinelli AR, Hamilton KA, Kauer JS. 1995. Salamander olfactory bulb neuronal activity observed by video rate, voltage-sensitive dye imaging. III. Spatial and temporal properties of responses evoked by odorant stimulation. J Neurophysiol 73:2053-2071.

Cometto-Muñiz JE, Cain WS, Abraham MH. 1998. Nasal pungency and odor of homologous aldehydes and carboxylic acids. Exp Brain Res 118:180-188.

Coopersmith R, Leon M. 1984. Enhanced neural response to familiar olfactory cues. Science 225:849-851.

Döving KB. 1966. An electrophysiological study of odour similarities of homologous substances. J Physiol 186:97-109.

Dravnieks A. 1985. Atlas of odor character profiles. Philadelphia: ASTM Data Series DS 61.

Friedrich RW, Korsching SI. 1997. Combinatorial and chemotopic odorant coding in the zebrafish olfactory bulb visualized by optical imaging. Neuron 18:737-752.

Galizia CG, Sachse S, Rappert A, Menzel R. 1999. The glomerular code for odor representation is species specific in the honeybee Apis mellifera. Nature Neurosci 2:473-478.

Gross-Isseroff R, Lancet D. 1988. Concentration-dependent changes of perceived odor quality. Chem Senses 13:191-204.

Guthrie KM, Gall CM. 1995. Functional mapping of odor-activated neurons in the olfactory bulb. Chem Senses 20:271-282.

Guthrie KM, Anderson AJ, Leon M, Gall C. 1993. Odor-induced increases in c-fos mRNA expression reveal an anatomical “unit” for odor processing in olfactory bulb. Proc Natl Acad Sci USA 90:3329-3333.

Hass HB, Newton RF. 1975. Correction of boiling points to standard pressure. In: Weast RC, editor. Handbook of chemistry and physics, 56th ed. Cleveland: CRC Press. p D176-D177.

Imamura K, Mataga N, Mori K. 1992. Coding of odor molecules by mitral/tufted cells in rabbit olfactory bulb. I. Aliphatic compounds. J Neurophysiol 68:1986-2002.

Joerges J, Küttner A, Galizia CG, Menzel R. 1997. Representations of odours and odour mixtures visualized in the honeybee brain. Nature 387:285-288.

Johnson BA, Leon M. 1996. Spatial distribution of [¹⁴C]2-deoxyglucose uptake in the glomerular layer of the rat olfactory bulb following early odor preference learning. J Comp Neurol 376:557-566.

Johnson BA, Woo CC, Duong H, Nguyen V, Leon M. 1995. A learned odor evokes an enhanced Fos-like glomerular response in the olfactory bulb of young rats. Brain Res 699:192-200.

Johnson BA, Woo CC, Leon M. 1998. Spatial coding of odorant features in the glomerular layer of the rat olfactory bulb. J Comp Neurol 393:457-471.

Johnson BA, Woo CC, Hingco EE, Pham KL, Leon M. 1999. Multidimensional chemotopic responses to n-aliphatic acid odorants in the rat olfactory bulb. J Comp Neurol 409:529-548.

Jourdan F, Duveau A, Astic L, Holley A. 1980. Spatial distribution of [¹⁴C]2-deoxyglucose uptake in the olfactory bulbs of rats stimulated with two different odours. Brain Res 188:139-154.

Kauer JS, Cinelli AR 1993. Are there structural and functional modules in the vertebrate olfactory bulb? Microsc Res Tech Feb 1:157-167

LaMotte RH, Mountcastle VB. 1975. Capacities of humans and monkeys to discriminate between vibratory stimuli of different frequency and amplitude: A correlation between neural events and psychophysical measurements. J Neurophysiol 38:539-559.

Lancet D. 1986. Vertebrate olfactory reception. Annu Rev Neurosci 9:329-355.

Laska M, Teubner P. 1998. Odor structure-activity relationships of carboxylic acids correspond between squirrel monkeys and humans. Am J Physiol 274:R1639-R1645.

Laska M, Galizia CG, Giurfa M, Menzel R. 1999. Olfactory discrimination ability and odor structure-activity relationships in honeybees. Chem Senses 24:429-438.

Malnic B, Hirono J, Sato T, Buck LB. 1999. Combinatorial receptor codes for odors. Cell 96:713-723.

Mombaerts P, Wang F, Dulac C, Chao SK, Nemes A, Mendelsohn M, Edmonson J, Axel R. 1996. Visualizing an olfactory sensory map. Cell 87:675-686.

Moncrieff RW. 1967. The Chemical Senses. London: Leonard Hill.

Mori K. 1987. Membrane and synaptic properties of identified neurons in the olfactory bulb. Progr Neurobiol 29:275-320.

Mori K, Mataga N, Imamura K. 1992. Differential specificities of single mitral cells in rabbit olfactory bulb for a homologous series of fatty acid odor molecules. J Neurophysiol 67:786-789.

Mori K, Yoshihara Y. 1995. Molecular recognition and olfactory processing in the mammalian olfactory system. Prog Neurobiol. 45:585-619.

O’Connell RJ, Stevens DA, Zogby LM. 1994. Individual differences in the perceived intensity and quality of specific odors following self- and cross-adaptation. Chem Senses 19:197-208.

Onoda N. 1992. Odor-induced fos-like immunoreactivity in the rat olfactory bulb. Neurosci Lett 137:157-160.

Pause BM, Sojka B, Ferstl R. 1997. Central processing of odor concentration is a temporal phenomenon as revealed by chemosensory event-related potentials (CSERP). Chem Senses 22:9-26.

Pilpel Y, Lancet D. 1999. The variable and conserved interfaces of modeled olfactory receptor proteins. Protein Sci 8:969-977

Polak EH. 1973. Multiple profile-multiple receptor site model for vertebrate olfaction. J Theor Biol 40:469-484.

Purdy D. 1931. Spectral hue as a function of intensity. Am J Psychol 43:541-559.

Ressler KJ, Sullivan SL, Buck LB. 1994. Information coding in the olfactory system: Evidence for a stereotyped and highly organized epitope map in the olfactory bulb. Cell 79:1245-1255.

Royet JP, Sicard G, Souchier C, Jourdan F. 1987. Specificity of spatial patterns of glomerular activation in the mouse olfactory bulb: Computer-assisted image analysis of 2-deoxyglucose autoradiograms. Brain Res 417:1-11.

Rubin BD, Katz LC. 1999. Optical imaging of odorant representations in the mammalian olfactory bulb. Neuron 23:499-511.

Sallaz M, Jourdan F. 1993. C-fos expression and 2-deoxyglucose uptake in the olfactory bulb of odour-stimulated awake rats. NeuroReport 4:55-58.

Sato T, Hirono J, Tonoike M, Takebayashi M. 1994. Tuning specificities to aliphatic odorants in mouse olfactory receptor neurons and their local distribution. J Neurophysiol 72:2980-2989.

Singer MS, Shepherd GM. 1994. Molecular modeling of ligand-receptor interactions in the OR5 olfactory receptor. Neuroreport. 5:1297-1300.

Singer MS, Oliveira L, Vriend G, Shepherd GM. 1995. Potential ligand-binding residues in rat olfactory receptors identified by correlated mutation analysis. Receptors Channels 3:89-95.

Singer MS, Weisinger-Lewin Y, Lancet D, Shepherd GM. 1996. Positive selection moments identify potential functional residues in human olfactory receptors. Receptors Channels 4:141-147.

Singer MS, Hughes TE, Shepherd GM, Greer CA. 1998. Identification of olfactory receptor mRNA sequences from the rat olfactory bulb glomerular layer. Neuroreport 9:3745-3748.

Slotnick BM, Panhuber H, Bell GA, Laing DG. 1989. Odor-induced metabolic activity in the olfactory bulb of rats trained to detect propionic acid vapor. Brain Res 500:161-168.

Snow WB. 1936. Change of pitch with loudness at low frequencies. J Acous Soc Am 8:14-19.

Stevens SS. 1935. The relation of pitch to intensity. J Acous Soc Am 6:150-154.

Stewart WB, Kauer JS, Shepherd GM. 1979. Functional organization of rat olfactory bulb analysed by the 2-deoxyglucose method. J Comp Neurol 185:715-734.

Theimer ET,Yoshida T, Klaiber EM. 1977. Olfaction and molecular shape. Chirality as a requisite for odor. J Agric Food Chem 25:1168-1177.

Tsuboi A, Yoshihara S, Yamazaki N, Kasai H, Asai-Tsuboi H, Komatsu M, Serizawa S, Ishii T, Matsuda Y, Nagawa F, Sakano H. 1999. Olfactory neurons expressing closely linked and homologous odorant receptor genes tend to project their axons to neighboring glomeruli on the olfactory bulb. J Neurosci 19:8409-8418.

Vassar R, Chao SK, Sitcheran R, Nuñez JM, Vosshall LB, Axel R. 1994. Topographic organization of sensory projections to the olfactory bulb. Cell 79:981-991.

Wang F, Nemes A, Mendelsohn M, Axel R. 1998. Odorant receptors govern the formation of a precise topographic map. Cell 93:47-60.

Yang X, Renken R, Hyder F, Siddeek M, Greer CA, Shepherd GM, Shulman RG. 1998. Dynamic mapping at the laminar level of odor-elicited responses in rat olfactory bulb by functional MRI. Proc Natl Acad Sci USA 95:7715-7720.

Yokoi M, Mori K, Nakanishi S. 1995. Refinement of odor molecule tuning by dendrodendritic synaptic inhibition in the olfactory bulb. Proc Natl Acad Sci USA 92:3371-3375.

Youngentob SL, Mozell MM, Sheehe PR, Hornung DE. 1987. A quantitative analysis of sniffing strategies in rats performing odor detection tasks. Physiol Behav 41:59-69.

Youngentob SL, Kent PF, Sheehe PR, Schwob JE, Tzoumaka E. 1995. Mucosal inherent activity patterns in the rat: Evidence from voltage-sensitive dyes. J Neurophysiol 73:387-398.

Table 1. Odorants used in this study.

IUPAC Common Chemical Functional

Name Name Structure Group

pentanoic acid valeric acid CH₃(CH₂)₃COOH carboxylic acid

methyl pentanoate methyl valerate CH₃(CH₂)₃COOCH₃ ester

1-pentanol amyl alcohol CH₃(CH₂)₃CH₂OH alcohol

pentanal valeraldehyde CH₃(CH₂)₃CHO aldehyde

2-hexanone methyl butyl ketone CH₃(CH₂)₃COCH₃ ketone

Table 2. Statistical significance of changes in centroids with odorant concentration.

Statistical degrees of

Odorant test freedom sector value P

pentanoic acid ANOVA 3, 16 lateral F = 0.29 n.s.*

medial F = 0.68 n.s.

methyl pentanoate t-test 4 lateral T = 0.67 n.s.

medial T = 0.94 n.s.

pentanol ANOVA 2, 6 lateral F = 1.73 n.s.

medial F = 0.90 n.s.

pentanal ANOVA 2, 6 lateral F = 3.03 n.s.

medial F = 6.44 < 0.05

2-hexanone ANOVA 2, 6 lateral F = 6.12 < 0.05

medial F = 21.43 < 0.01

* n.s., not significant

Figure legends

Figure 1. Maps of 2-DG uptake across the entire glomerular layer in rats exposed to different concentrations of five chemically related odorants. Measurements of 2-DG uptake were taken at fixed angle increments in coronal sections spaced evenly throughout the rostral-caudal extent of the olfactory bulb. These values (about 2500 per bulb) were organized into arrays of angle increment x section number. After correction for differences in size, two arrays representing the two bulbs of each animal were averaged. Average arrays from animals exposed to air only then were subtracted, and each resultant array was transformed to an array of z scores relative to the mean and standard deviations of all values across the array. Arrays from all animals exposed to the same concentration of the same odorant were then averaged together (5 rats were exposed to each concentration of pentanoic acid, 3 to each concentration of the other odorants). The arrays were visualized as contour charts using Microsoft Excel 98. The resulting maps thus depict the glomerular layer as a two-dimensional surface opened along the dorsal extremity of the bulb (upper left panel). Lateral is present in the upper half of each map, and rostral is to the left. Z score values are color-coded such that high values are assigned warm colors and low values are assigned cool colors. Asterisks indicate odorant concentrations that were found to give patterns distinct from those present in air-exposed control animals as determined in subsequent analyses. For pentanal and 2-hexanone, black arrows are used to denote parts of the spatial pattern that were consistent in location across all concentrations giving an odorant-evoked pattern of activity and that became minor parts of the patterns at higher odorant concentrations. White arrows indicate areas of response detected at higher, but not lower concentrations of the two odorants. Scale bar = 2 mm.

Figure 2. Glomerular modules were defined by comparing simple odorant-evoked z score patterns with increasingly complex ones (Fig.1). Each time a region of response unique to an odorant or concentration was detected in the lateral aspect of the bulb, a new area of activity also was detected in the medial aspect. To illustrate this lateral/medial symmetry, lower case letters are used to label the lateral modules, and corresponding upper case letters are used to label the medial modules. Module I of the medial bulb had no lateral equivalent.

Figure 3. Amounts of 2-DG upta