Emergence of dual coding mechanisms in a network model of the locust antennal lobe: Transient temporal binding vs. a high dimensional rate code

The antennal lobe (AL) is the primary structure within the locust's brain that receives information from the olfactory receptor neurons (ORNs) within the antennae that detect environmental odor signals. Different odors activate distinct subsets of ORNs, implying that neuronal signals at the level of the antennae encode odors in a combinatorial fashion. Within the AL, however, different odors produce signals with long-lasting dynamic transients carried by overlapping neural ensembles, suggesting a more complex coding scheme. In this work, I used a large-scale point neuron model of the locust AL to investigate this shift in stimulus encoding and potential consequences for odor discrimination. In accordance with experiment, the model produces stimulus-sensitive, dynamically evolving populations of active AL neurons. The model relies critically on the persistence time-scale associated with ORN input to the AL, sparse PN-PN connectivity, weak excitatory coupling, and a synaptic slow inhibitory mechanism. Immediately after stimulus onset, fast inhibition induces higher order correlations across PNs that give rise to a neural code based on temporal binding, a code which is subsequently disrupted by the activation of a slow inhibitory current. Following the disappearance of the temporal code, a rate code emerges in which architectural and dynamical network features collectively generate odor representations of higher dimension than the stimulus space, leading to more uniform stimulus separation and improved performance by linear classifiers.