Vector coordinate transformation in the Drosophila navigation system

Abstract

Animals navigational behaviors rely on internal representations of space, yet how these representations guide behaviors is poorly understood. Here, we determined how neurons in the brain of the fruit fly, Drosophila melanogaster, transform a world-centric (allocentric) reference frame to one that is body-centric (egocentric), which is essential for navigating to a goal. PFL1 neurons in the flys central complex, the navigation center, have two dendrites; one of these receives input from neurons that encode a head direction vector, while the other receives input from neurons in a neuropil called the fan-shaped body (FB). Therefore, we investigated whether PFL1 neurons transform allocentric vectors to an egocentric reference frame. We developed a computational model that shows how PFL1 can transform an allocentric direction vector encoded by FB neurons to an egocentric reference frame by nonlinearly combining it with the head direction input. In tethered flying flies, we simultaneously recorded the head direction input to as well as the output of PFL1, which we used to constrain and fit our model. This allowed us to infer that PFL1 neurons receive a stored direction that persists across behavioral trials for the same fly but appears to be random across flies, fundamental characteristics of a goal. Our results describe how a population of neurons performs a coordinate transformation that can facilitate the pursuit of a goal, a fundamental computation for any neural system containing representations in multiple reference frames.