Motion-based Spatial Vision in Drosophila: Connecting Canonical Levels of Understanding in Neuroscience

Date of Award

Spring 2022

Document Type


Degree Name

Doctor of Philosophy (PhD)


Interdepartmental Neuroscience Program

First Advisor

Clark, Damon


A canonical metatheoretical framework in neuroscience states that, for one to understand a brain, parallel descriptions at the levels of computational theory, algorithms, and implementations, are necessary. However, computational goals of a brain are less empirically accessible than its algorithms or mechanisms, making analyses at the theoretical level challenging. The present thesis proposes that studying spatial vision problems in fruit fly Drosophila can be an interesting model case where one can bridge the computation theoretical and the other levels, aided by simple geometry of spatial vision problems as well as sophisticated circuit dissection tools in Drosophila. Chapter 2 explores how walking flies avoid collisions with other animals by using back-to-front motion as a heuristic cue for imminent collisions. With behavioral and two-photon calcium imaging experiments, I demonstrate that neurons called LPLC1 implements the collision avoidance behavior, and show that their visual tuning mirrors the geometry of collisions. In addition, I explore circuitry surrounding LPLC1 with genetic and connectomic tools to understand how tuning for collision cues is achieved. Chapter 3 demonstrates how flies can adaptively suppress their course-stabilizing optomotor response in the presence of stationary visual patterns, a heuristic visual cue indicating the absence of self-rotation. With calcium imaging of the motion detection pathway as well as behavioral genetic screening, I constrain the circuit that detects stationary patterns to suppress optomotor response. In the last chapter, I will discuss how well the two projects achieved the goal of bridging the canonical levels of understanding, as well as possible strategies to expand the scope of the functionalist framework beyond simple problems of spatial vision.

This document is currently not available here.