Distinct circuit motifs evaluate opposing innate values of odors

Abstract

Evaluating the innate value of objects is critical for expressing adaptive behaviors. However, where and how this computation takes place in the brain remain elusive. By recording from virtually every neuron in Drosophila higher olfactory areas, we show that the lateral horn is a site of innate odor value computation, where distinct neurons represent opposing innate values. A connectome-based spiking network model recapitulating the neural activity indicates that representations of aversive odors emerge through specific convergence of feedforward excitation, whereas those of attractive odors emerge through additional local inhibition. This inhibition is broad yet balanced with excitation and implements gain control and thresholding to shape attractive odor tuning. Manipulation of local inhibition biased neuronal and behavioral odor responses according to the prediction of the model. Thus, odors at the opposite ends of the hedonic spectrum are processed in sub-circuits that are not only segregated but also distinct in connectivity motifs.

Keywords: Ca(2+) imaging; Drosophila; circuit modeling; connectome; innate value; lateral horn; olfaction.