The field and behavioral studies of M. sexta's olfactory-mediated behaviors to D.
wrightii and A. palmeri flowers provided an ecological context to discover the neural
circuits underlying these plant-pollinator interactions. I use an interdisciplinary
approach by coupling odor quantification (gas chromatography with mass spectrometric
detection), behavioral wind tunnel experiments, and electrophysiology, to determine
how the complex floral odors, at natural concentrations, are encoded in the moth
antennal lobe (AL). Use of a tandem gas chromatography-multi-channel recording system
(GC-MR) allowed determination of the key single odorants within the complex floral
mixtures that AL neurons are especially responsive to (Riffell et al., Current Biology
19: 335-340). These results demonstrate that neural ensemble responses to the two
floral mixtures can be reproduced by a much smaller subset of odorants (<5). In turn,
behavioral wind tunnel results demonstrate these mixtures containing the smaller
subset of odorants are as effective as the natural floral mixtures (Lei et al., J.
Biol. 8: 1-16). Finally, spatiotemporal encoding by the AL neural ensemble provides
a means by which the olfactory system can generalize between related mixtures while
providing enough contrast for discrimination between floral odors (Riffell et al.,
PNAS In press).
To determine how the antennal lobe (AL) of Manduca sexta encodes behaviorally relevant floral mixtures and single constituents, a multi-channel neural-ensemble recording array was coupled with a GC-FID and GC-MS. Integration of these technologies allows examination of AL response to natural complex mixtures and provides a means of fractionating those same blends into their single components. (Top) The multiunit recordings were conducted in the AL, which is the first processing center of complex olfactory information in the insect central nervous system. With multiunit extracellular electrodes, distinct locations within the CNS can be simultaneously recorded from, thereby allowing spatial and temporal comparisons in the neural network responses to complex odors. (Immediate left) Neuron responses to the D. wrightii headspace fractionated blend components from the GC (Bottom trace in red) revealed units that were distinctly tuned to one or two components of the natural blend (Middle trace in blue), or were broadly active to multiple components (Top trace in green).
