Olfactory Coding: Time in a Model

stimulus. In rats, however, comparatively large OB leIn recent years, there have been dramatic advances sions, even lesions that include sites most activated by in olfactory research, with important implications for the test odorant, had little effect on performance of understanding neural coding in this system. Yet we still several olfactory tasks requiring odor detection and rechave a surprisingly poor understanding of how odor ognition (Slotnick et al., 1997). These results suggest information is encoded, even at the earliest stages of that activation of specific areas of the OB is not necescentral olfactory processing. This is partly because of sarily a critical component of the olfactory code. It may the paucity of reported studies that go beyond descripbe that local interactions and relative activity levels tion of the system to test hypotheses about specific across the OB are more informative than absolute levels mechanisms of olfactory coding. The recently published at specific sites. In an interesting set of experiments, paper by Stopfer et al. (1997) is remarkable, both beMori and his colleagues have begun to examine how cause it is one of those very fewdirect tests of a hypothespatial relationships in OB input give rise to patterns of sis on coding in olfaction and because it reports a correresponsivity characteristic of the OB output neurons, lation between a disruption in temporal characteristics the mitral and tufted cells (MTs; for a review of these of neural response and a specific and limited modulation experiments, see Mori and Yoshihara, 1995). Individual of an odor-guided behavior. Their work opens the door MTs respond with excitation to a set of similarly structo a line of research testing the contribution of action tured compounds and are likely to be suppressed by potential timing and synchronization to coding of senodorants whose chemical structures lie just outside that sory information in the nervous system. set (Yokoi et al., 1995). Blocking inhibitory input to MTs Much of the current thought on olfactory coding foeliminates the suppressive but not the excitatory recuses on the importance of spatial, rather than temporal, sponses. Inhibitory connections between neighboring patterns of activity (for review, see Mori and Yoshihara, glomeruli receiving similar inputs appear to account for 1995; Buck, 1996). The concept of an orderly mapping the suppressive responses, analogous to the lateral inbetween stimulus dimensions and neuronal position, as hibitory basis for center/surround receptive fields in the found in other sensory systems, is compelling and may visual system (Yokoi, et al., 1995). in some sense apply to olfactory systems as well. At Though the contribution of spatial patterns to olfacthe level of the primary central target of the olfactory tory coding has received the majority of the attention receptor neurons (in vertebrates, the olfactory bulb of late, temporal aspects of neuronal response have [OB]), electrophysiological recordings, c-fos and 2-deoxylong been considered a crucial component of an olfacglucose experiments, and optical recordings have all tory code (see, for example, Adrian, 1953). Temporal shown that spatial patterns of activity differ for different characteristics of response have been described in the olfactory stimuli (for example, Joerges et al., 1997; also patterns of single unit spiking and in oscillatory activity see references in Mori and Yoshihara, 1995). Molecular of populations of neurons in olfactory processing areas. biological data on expression patterns of the genes that In individual MTs of the vertebrate OB, temporal patterns are believed to encode olfactory receptor proteins indithat consist of periods of spiking and suppression vary cate an underlying basis for those patterns: each of the within the same cell for different odorant stimuli and perhaps hundreds of receptor neuron types, as characchange nonmonotonically with changes in concentraterized by the one or few kinds of receptor genes they tion of the same stimulus (see Hamilton and Kauer, 1989, express, converges at specified locations of the OB and references therein). Similar patterns of output cell (Mombaerts, 1996). The fact that these locations vary spiking have been described in invertebrates as well little among individuals of a species has strengthened (Laurent et al., 1996). Though it is hypothesized that the idea that the OB includes a functional map, or represuch temporal patterns carry informationabout stimulus sentation, of olfactory receptor space (for references, quality and intensity, their precise role in olfactory codsee Buck, 1996). Experiments using optical recording ing is unknown. Likewise, there are various hypotheses during olfactory stimulation in zebrafish have begun to about the contribution of population oscillation and demonstrate correlations between physical–chemical waves of activity to olfactory coding. Many animals,