Contrast enhancement of stimulus intermittency in a primary olfactory network and its behavioral significance

Abstract

Background: An animal navigating to an unseen odor source must accurately resolve the spatiotemporal distribution of that stimulus in order to express appropriate upwind flight behavior. Intermittency of natural odor plumes, caused by air turbulence, is critically important for many insects, including the hawkmoth, Manduca sexta, for odor-modulated search behavior to an odor source. When a moth's antennae receive intermittent odor stimulation, the projection neurons (PNs) in the primary olfactory centers (the antennal lobes), which are analogous to the olfactory bulbs of vertebrates, generate discrete bursts of action potentials separated by periods of inhibition, suggesting that the PNs may use the binary burst/non-burst neural patterns to resolve and enhance the intermittency of the stimulus encountered in the odor plume.

Results: We tested this hypothesis first by establishing that bicuculline methiodide reliably and reversibly disrupted the ability of PNs to produce bursting response patterns. Behavioral studies, in turn, demonstrated that after injecting this drug into the antennal lobe at the effective concentration used in the physiological experiments animals could no longer efficiently locate the odor source, even though they had detected the odor signal.

Conclusions: Our results establish a direct link between the bursting response pattern of PNs and the odor-tracking behavior of the moth, demonstrating the behavioral significance of resolving the dynamics of a natural odor stimulus in antennal lobe circuits.

Figure 1

Effects of bicuculline on the firing pattern of MGC-PNs. (a) Shown as raw spike traces, bath application of 100 μM bicuculline changed the spontaneous firing pattern of an MGC-PN from a random bursting (left) to a more regular tonic pattern (middle). This change was reversed with saline wash (right). (b) The M inhibitory period (I₂) that typically follows the odor-evoked excitatory phase in MGC-PNs (upper panel) was completely blocked by treatment with 200 μbicuculline, resulting in an extended excitatory response (asterisks, lower panel). Odor pulse is indicated by the black bar below the traces. (c) Confocal micrograph showing the lucifer yellow fluorescent mark (arrowed) in the cumulus (C) deposited by the glass electrode used to record the C-PN in (b). T, toroid I. (d) Graphs of peristimulus responses (derived from five odor pulses) of 25 MGC-PNs to their specific ligands under saline control (blue curve; mean ± SEM) and bicuculline treatment (orange curve; mean ± SEM) at low (25 μM, n = 8), intermediate (50 μM or 100 μM, n = 7), and high (200–500 μM, n = 10) dosages. The onset of the 50 ms stimulus was at time zero. (e) Histograms derived from the graphs in (d). The shaded areas represent the I₂ period, during which the averaged firing rate was not significantly different (NS) between low-dose bicuculline treatment and saline control, but was significantly elevated by intermediate and high-dosage bicuculline treatment. The abbreviation ns and the asterisks respectively indicate non-statistical (Mann Whitney U test, p > 0.05 for low dose, n = 8) and statistical significance (Mann Whitney U test, p < 0.03 for intermediate dose, n = 7; p < 0.001 for high dose, n = 10).

Figure 2

Bicuculline-effects on PNs' pulse-tracking capability. (a) Autocorrelation-based pulse-following index (PFI) was calculated to quantify the capability of PNs to track odor pulses delivered at 1 Hz repetition rate under saline control (left), bicuculline treatment (center), and saline wash (right). The raster plots above the correlograms illustrate the response of a T-PN to two consecutive odor pulses. Note that the drop in PFI value during bicuculline treatment is consistent with the decreased pulse resolution shown in the raster plots. (b-e) Population data (mean ± SEM) showing that bicuculline treatment consistently decreases the PFI values. (b, c) This effect was independent of stimulus type: (b) blend; (c) individual excitatory stimulus component. However, the PFI profiles for (d) T-PNs and (e) C-PNs were dramatically different, with C-PNs having higher PFI values in the range 0.2–1 Hz than the T-PNs under saline control (solid line), thus resulting in a greater drop in PFI values from control to bicuculline treatment (dotted line). Asterisks indicate statistical significance between control and drug treatment (repeated-measure two-way ANOVA at p = 0.05 level). (f-g) Stacked correlograms derived from the responses of ten PNs to their specific ligands show their capability to track odor pulses delivered at various repetition rates (ranging from 0.2 to 10 Hz) under (f) saline control and (g) bicuculline treatment. The pseudocolor scale, indicating the correlation coefficient, applies to both panels.

Figure 3

Bicuculline significantly affects pheromone-mediated navigation behavior. (a-c) Behavioral measurements on unoperated (gold), saline-injected (cyan) and bicuculline-injected (red) moths in a wind tunnel supplied with (a) pheromone or (b) solvent control (cyclohexane). Neither bicuculline nor saline injection affected a moth's ability to be motivated to fly (wing-fanning) or make upwind progress. A significantly lower percentage of bicuculline-injected moths (n = 12) displayed close hover, source contact and abdomen curl, compared with the unoperated (n = 10) and saline-injected (n = 15) groups (G test: p < 0.05). Under cyclohexane, all moths showed wing-fanning behavior, but only 30–50% of moths in each group (n = 10, 6, 9 for unoperated, saline-injected and bicuculline-injected, respectively) progressed upwind and an even lower percentage displayed close hover and source contact. None of the animals that came close to the source displayed abdomen curl. (c) The effects of bicuculline on close hover, source contact and abdomen curl shown in (a) were reversed after recovery for at least 2 h in a dark environmental chamber (n = 8, 7, 9 for unoperated, saline-injected and bicuculline-injected, respectively). Different letters within a behavioral category denote statistical significance (G test: p <0.05). (d-i) Flight-track analysis on unoperated (d, g), saline-injected (e, h) and bicuculline-injected (f, i) moths with pheromone or solvent control in the wind tunnel. (d, e) Using pheromone as the odor source, the unoperated and saline-injected moths flew directly toward the odor source, thus resulting in approximately straight flight tracks (top), centrally distributed transit probability (middle panels) and track-angle distribution histograms (bottom panels) with a prominent peak at zero degrees (mean ± SEM). The central distribution of transit probability is further demonstrated with a summed bar graph (along the wind direction) located to the right of the pseudocolor plots, showing a single peak at the center. (f) Bicuculline-injected moths, on the other hand, markedly diminished the central peak as well as the tracking frequency peak at zero degree track angle. (g-i) Replacing the pheromone with solvent control (cyclohexane) in the wind tunnel resulted in unanimous 'looping' flight tracks in all three treatment groups, reflecting an engagement of cross-wind casting in these moths, which is also shown in the randomly distributed transit probability of occupancy as well as in the bimodal distribution of track angle histograms.

Figure 4

Injection of bicuculline into the MGC does not influence a moth's abilities to navigate to floral odors. (a) Behavioral measurements on unoperated (purple), saline-injected (blue) and bicuculline-injected (green) moths in a wind tunnel supplied with a floral odor. For all behaviors, there were no significant differences between treatments (G test: p > 0.10). n = 3–8 moths per treatment. (b) Measurement of track angles of bicuculline-injected moths flying toward floral odor source. A prominent peak at zero degrees indicates that the drug injected into MGC did not affect their navigation behavior mediated by floral odor. (c) Moth flight tracks to pheromone (orange) and floral odors (green, blue and violet). When injected into the MGC, bicuculline caused moths to increase the number of casts in the flight and a decrease in the ability to locate the pheromone source (orange flight tracks). In contrast, bicuculline injected into the MGC did not influence the ability of the moths to successfully navigate to, and locate, the floral odor source (green flight tracks). Saline-injected (blue flight tracks) and unoperated (violet flight tracks) moths exhibited similar flight behaviors to the floral odor as those moths treated with bicuculline. For each treatment three moth flight tracks were selected using a random number generator (denoted by tracks of different color shades). The tracks are made up of circles corresponding to video images captured at 0.016 s intervals.