Elements of Olfactory Intelligence in Drosophila

Abstract

The ability to make the world of odorants intelligible is a key capability of the Drosophila olfactory system that we shall call olfactory intelligence. Characterizing the functional logic of the Drosophila early olfactory system that makes the natural world of odorants intelligible is a major challenge in neuroscience. Starting by modeling the space of odorants using constructs of both semantic and syntactic information, we establish that the Antenna Lobe and Mushroom Body Calyx first decompose the confounding representation of the Antenna into a concentration independent odorant semantics information and the ON-OFF timing of the syntactic information. Subsequently the two streams of information are integrated and produce a novel time and rank-based representation of Kenyon Cell outputs, called the marked first spike sequence code. In conjunction with a novel distance measure for the spike sequence code, we demonstrate that the rank-based representation supports accurate classification of ON-OFF odorant semantics. Computationally, these elements of olfactory intelligence are realized by a class of differential divisive normalization processors modeling the feedback circuits in a causal chain of stages including the Antenna, Antennal Lobe and the Mushroom Body Calyx. Consequently, the early olfactory system makes the natural world of odorants intelligible.