Excitatory Interactions between Olfactory Processing Channels in the Drosophila Antennal Lobe

Biological Mechanisms for Learning: A Computational Model of Olfactory Learning in the Manduca sexta Moth, With Applications to Neural Nets

The insect olfactory system, which includes the antennal lobe (AL), mushroom body (MB), and ancillary structures, is a relatively simple neural system capable of learning. Its structural features, which are widespread in biological neural systems, process olfactory stimuli through a cascade of networks where large dimension shifts occur from stage to stage and where sparsity and randomness play a critical role in coding. Learning is partly enabled by a neuromodulatory reward mechanism of octopamine stimulation of the AL, whose increased activity induces synaptic weight updates in the MB through Hebbian plasticity. Enforced sparsity in the MB focuses Hebbian growth on neurons that are the most important for the representation of the learned odor. Based upon current biophysical knowledge, we have constructed an end-to-end computational firing-rate model of the Manduca sexta moth olfactory system which includes the interaction of the AL and MB under octopamine stimulation. Our model is able to robustly learn new odors, and neural firing rates in our simulations match the statistical features of in vivo firing rate data. From a biological perspective, the model provides a valuable tool for examining the role of neuromodulators, like octopamine, in learning, and gives insight into critical interactions between sparsity, Hebbian growth, and stimulation during learning. Our simulations also inform predictions about structural details of the olfactory system that are not currently well-characterized. From a machine learning perspective, the model yields bio-inspired mechanisms that are potentially useful in constructing neural nets for rapid learning from very few samples. These mechanisms include high-noise layers, sparse layers as noise filters, and a biologically-plausible optimization method to train the network based on octopamine stimulation, sparse layers, and Hebbian growth.

A Population of Projection Neurons that Inhibits the Lateral Horn but Excites the Antennal Lobe through Chemical Synapses in Drosophila

In the insect olfactory system, odor information is transferred from the antennal lobe (AL) to higher brain areas by projection neurons (PNs) in multiple AL tracts (ALTs). In several species, one of the ALTs, the mediolateral ALT (mlALT), contains some GABAergic PNs; in the Drosophila brain, the great majority of ventral PNs (vPNs) are GABAergic and project through this tract to the lateral horn (LH). Most excitatory PNs (ePNs), project through the medial ALT (mALT) to the mushroom body (MB) and the LH. Recent studies have shown that GABAergic vPNs play inhibitory roles at their axon terminals in the LH. However, little is known about the properties and functions of vPNs at their dendritic branches in the AL. Here, we used optogenetic and patch clamp techniques to investigate the functional roles of vPNs in the AL. Surprisingly, our results show that specific activation of vPNs reliably elicits strong excitatory postsynaptic potentials (EPSPs) in ePNs. Moreover, the connections between vPNs and ePNs are mediated by direct chemical synapses. Neither pulses of GABA, nor pharmagological, or genetic blockade of GABAergic transmission gave results consistent with the involvement of GABA in vPN-ePN excitatory transmission. These unexpected results suggest new roles for the vPN population in olfactory information processing.

Space Takes Time: Concentration Dependent Output Codes from Primary Olfactory Networks Rapidly Provide Additional Information at Defined Discrimination Thresholds

As odor concentration increases, primary olfactory network representations expand in spatial distribution, temporal complexity and duration. However, the direct relationship between concentration dependent odor representations and the psychophysical thresholds of detection and discrimination is poorly understood. This relationship is absolutely critical as thresholds signify transition points whereby representations become meaningful to the organism. Here, we matched stimulus protocols for psychophysical assays and intracellular recordings of antennal lobe (AL) projection neurons (PNs) in the moth Manduca sexta to directly compare psychophysical thresholds and the output representations they elicit. We first behaviorally identified odor detection and discrimination thresholds across an odor dilution series for a panel of structurally similar odors. We then characterized spatiotemporal spiking patterns across a population of individually filled and identified AL PNs in response to those odors at concentrations below, at, and above identified thresholds. Using spatial and spatiotemporal based analyses we observed that each stimulus produced unique representations, even at sub-threshold concentrations. Mean response latency did not decrease and the percent glomerular activation did not increase with concentration until undiluted odor. Furthermore, correlations between spatial patterns for odor decreased, but only significantly with undiluted odor. Using time-integrated Euclidean distance (ED) measures, we determined that added spatiotemporal information was present at the discrimination but not detection threshold. This added information was evidenced by an increase in integrated distance between the sub-detection and discrimination threshold concentrations (of the same odor) that was not present in comparison of the sub-detection and detection threshold. After consideration of delays for information to reach the AL we find that it takes ~120-140 ms for the AL to output identity information. Overall, these results demonstrate that as odor concentration increases, added information about odor identity is embedded in the spatiotemporal representation at the discrimination threshold.

All-or-none population bursts temporally constrain surround inhibition between mouse olfactory glomeruli

With each sniff, the olfactory bulbs of the brain generate a neural activity pattern representing the odour environment, transmitting this to higher brain centres in the form of mitral cell output. Inhibitory circuits in the olfactory bulb glomerular and external plexiform layers may amplify contrast in these patterns, through surround inhibition of mitral cells. These circuits may operate in series, but their respective roles are unclear. A single sniff is sufficient for odour discrimination, but is not clear that the inhibitory circuits act within this timeframe. We used microdissected slices of mouse olfactory bulb to study each circuit in isolation. We found that unlike surround inhibition mediated in the external plexiform layer, surround inhibition mediated in the glomerular layer was activated by sensory synaptic input, but not by mitral cell output. The results also suggest that interactions between olfactory glomeruli are exclusively inhibitory, unlike in antennal lobe, and that surround inhibition mediated within the external plexiform layer may involve neural circuit elements not preserved in slice preparations. Surround inhibition was effective only after an interval corresponding to a single sniff in vivo. Surplus excitation, initiated by sensory input but generated by collective all-or-none responses of mitral cells, may delay surround inhibition and allow the synchronous activation of multiple glomeruli without each suppressing the other. Surround inhibition in the glomerular layer may subsequently allow a fresh representation of the odour environment to be generated with each sniff. These findings are consistent with combinatorial odour coding based on all-or-none glomerular responses.