Descending premotor target tracking systems in flying insects Jack Alexander Supple Department of Physiology, Development, and Neuroscience University of Cambridge This dissertation is submitted for the degree of Doctor of Philosophy Darwin College September 2019 Declaration This dissertation is the result of my own work and includes nothing which is the outcome of work done in collaboration except as declared in the preface and specified in the text. It is not substantially the same as any that I have submitted, or, is being concurrently submitted for a degree or diploma or other qualification at the University of Cambridge or any other University or similar institution. I further state that no substantial part of my dissertation has already been submitted, or, is being concurrently submitted for any such degree, diploma or other qualification at the University of Cambridge or any other University or similar institution. This dissertation contains fewer than 60,000 words, exclusive of tables, footnotes, bibliography, and appendices. i Descending premotor target tracking systems in flying insects By Jack Alexander Supple Abstract The control of behaviour in all animals requires efficient transformation of sensory signals into the task-specific activation of muscles. Predation offers an advantageous model behaviour to study the computational organisation underlying sensorimotor control. Predators are optimised through diverse evolutionary arms races to outperform their prey in terms of sensorimotor coordination, leading to highly specialised anatomical adaptations and hunting behaviours, which are often innate and highly stereotyped. Predatory flying insects present an extreme example, performing complex visually-guided pursuits of small, often fast flying prey over extremely small timescales (Olberg et al., 2007; Wardill et al., 2015, 2017). Furthermore, this behaviour is controlled by a tiny nervous system, leading to pressure on neuronal organisation to be optimised for coding efficiency (Gonzalez-Bellido et al., 2011). In Dragonflies, a population of eight pairs of bilaterally symmetric Target Selective Descending Neurons (TSDNs) relay visual information about small moving objects from the brain to the thoracic motor centres. These neurons encode the movement of small moving objects across the dorsal fovea region of the eye which is fixated on prey during predatory pursuit, and are thought to constitute the commands necessary for actuating an interception flight path (Gonzalez-Bellido et al., 2013; Olberg, 1986; Olberg et al., 2007). TSDNs are characterised by their receptive fields, with responses of each TSDN type spatially confined to a specific portion of the dorsal fovea visual field and tuned to a specific direction of object motion (Gonzalez-Bellido et al., 2013). To date, little is known about the descending representations mediating target tracking in other insects. This dissertation presents a comparative report of descending neurons in a variety of flying insects. The results are organised into three chapters: Chapter 3 identifies TSDNs in demoiselle damselflies and compares their response properties to those previously described in dragonflies. Demoiselle TSDNs are also found to integrate binocular information, which is further elaborated with prism and eyepatch experiments. Chapter 4 describes TSDNs in two dipteran species, the robberfly Holcocephala fusca and the killerfly Coenosia attenuata. ii Chapter 5 describes an interaction between small- and wide-field visual features in TSDNs of both predatory and nonpredatory dipterans, finding functional similarity of these neurons for prey capture and conspecific pursuit. Dipteran TSDN responses are repressed by background motion in a direction dependent manner, suggesting a control architecture in which target tracking and optomotor stabilization pathways operate in parallel during pursuit. iii Acknowledgements I am thankful to Dr Paloma Gonzalez-Bellido for welcoming me into her lab and all the help over the years. Thanks also to Dr Trevor Wardill for helping to get me up and running with experiments, and to all other members of the Fly Systems lab, especially Mary Sumner, Dr Kate Feller, Sergio Rossoni, and Sam Fabian. Special thanks to Daniel Pinto-Benito for being an excellent student and friend. It was a pleasure to collaborate with Dr Karin Nordström, and I am extremely grateful for her extensive and imaginative efforts. Thanks also to Professor Rob Olberg for his intracellular efforts in demoiselles, and for being such a pleasure to work with. Thanks to Professor Tom Daniel for hosting me in his lab and for lending inspiration throughout the years. Special thanks to Dr Kristian Franze for helping throughout the transition from Cambridge to Minnesota. Contributions A large part of this work results from the efforts of members of the Fly Systems lab and external collaborators. Chapters 3 and 4 include data collected in part by Paloma Gonzalez-Bellido, my supervisor; Daniel Pinto-Benito, a summer student under my mentorship; and more recently Siddhant Pusdekar, Molly Liu, and Daniel Geleano who collected 2019 Calopteryx maculata behavioural data. In each case I was responsible for experimental design and data analysis, in addition to collecting a large portion of the dataset myself. Chapter 3 also includes data collected and analysed by Professor Rob Olberg; I was responsible for processing anatomical data and synthesising the data. Chapter 5 is based on a paper published in collaboration with Dr Karin Nordström’s laboratory at Flinder’s University, Australia. Dr Nordström’s group was responsible for the experimental design, acquisition and analysis of all hoverfly data in the study. I was responsible for the experimental design, collection and analysis of all robberfly data, and contributed to the interpretation of the combined datasets. Other specific contributions throughout this thesis are documented in an ‘Acknowledgements and contributions’ section at the beginning of each results chapter. Jack Supple September 2019 iv Table of Contents Declaration i Abstract ii Acknowledgements iv Contributions iv Table of Contents v List of Figures x Chapter 1: Introduction 1 1.1 Target tracking and chasing 1 1.1.1 Definitions of biological sameness: homology and analogy 2 1.1.2 Diversity of chasing behaviours 4 1.2 The organisation of the insect visual system 7 1.2.1 Compound eyes 7 1.2.2 Ocelli 13 1.2.3 Brain organisation and nomenclature 13 1.2.4 The optic lobes 15 1.2.5 The central brain 17 1.2.6 Descending neurons and the ventral nerve cord 19 1.3 Neuronal mechanisms underlying motion vision 19 1.3.1 Elementary motion detection 19 1.3.2 Wide-field motion processing 23 1.3.2.1 Lobula Plate Tangential Cells (LPTCs) 23 1.3.2.2 Wide-field sensitive descending neurons 26 1.3.3 Target motion processing 28 1.3.3.1 Target selective lobula complex neurons 28 1.3.3.2 Target Selective Descending Neurons (TSDNs) 37 1.4 Models for visual control of target chasing 39 1.4.1 Land and Collett model 39 1.4.2 Inner- and outer-loop control 41 v 1.5 Aims and structure of the thesis 46 Chapter 2: Methods and Materials 49 2.1 Animals 50 2.1.1 Odonata 50 2.1.2 Diptera 50 2.2 Dissections 51 2.2.1 Ventral nerve cord 51 2.3 Visual stimuli 53 2.3.1 Using the DepthQ 360 DLP projector 53 2.3.2 StimGL 56 2.3.3 Tracking stimulus presentation 56 2.4 Electrophysiology 57 2.4.1 Extracellular recordings and spike sorting 57 2.4.2 Eyepatches and prisms 59 2.4.3 Data analysis 61 2.4.3.1 Visual receptive field mapping 61 2.5 Anatomy 67 2.5.1 2-Photon whole brain imaging 67 2.5.2 Macrophotography and pseudopupil measurements 68 2.6 Highspeed videography 71 2.6.1 Visual stimuli 71 2.6.2 Data acquisition 71 2.6.3 Calibration 73 2.6.4 Digitisation and 3D reconstruction 73 2.6.5 Behavioural analysis 75 Chapter 3: Binocular encoding in the premotor target tracking system of damselflies 77 3.1 Introduction 77 3.2 Methods and materials 79 vi 3.2.1 Animals 79 3.2.2 High-speed videography of predation 79 3.2.3 Electrophysiology 81 3.2.4 Visual stimuli 82 3.2.5 Pseudopupil measurements 82 3.3 Results 82 3.3.1 Odonate ocular morphology 82 3.3.2 Attack trajectory in damselflies 86 3.3.3 TSDNs serving the demoiselle frontal fovea 86 3.3.4 Demoiselle TSDNs are binocular, exhibiting binocular-only, ocular-balanced, or ocular-dominant responses 93 3.3.5 Differences in global light intensity do not underlie the binocular input requirements of TSDNs 96 3.3.6 Demoiselle TSDN receptive fields resulting from reduced binocularity are consistent with binocular summation 96 3.4 Discussion 101 3.4.1 Eye morphology, hunting strategy, and TSDN homology within Odonata 101 3.4.2 Neuronal encoding of holoptic versus dichoptic visual space 102 3.4.3 Binocular properties of demoiselle TSDNs 103 3.5 Supplementary information 106 Chapter 4: Target Selective Descending Neurons in convergent predatory dipterans 133 4.1 Introduction 113 4.2 Methods and materials 115 4.2.1 Animals and electrophysiology 115 4.2.2 2-Photon brain imaging 116 4.2.3 Data analysis 116 4.3 Results 116 4.3.1 Neuroanatomy and descending neuron organisation 116 vii 4.3.2 Dipteran Target Selective Descending Neurons (dTSDNs) in robberflies and killer flies 119 4.3.3 Holcocephala dTSDN receptive fields 124 4.3.4 Coenosia dTSDN receptive fields 126 4.4 Discussion 127 dTSDN response latencies 127 4.4.1 dTSDN receptive fields 128 4.4.2 Diversity of dipteran TSDNs 129 Chapter 5: Integration of small- and wide-field visual features in Target Selective Descending Neurons