Periodic Reporting for period 1 - MOBY-FLY (Circuit mechanisms of behavioural variability in Drosophila flight.)
Reporting period: 2022-12-01 to 2025-05-31
At any given moment, we are bombarded by a variety of different stimuli detected by our sensory systems and we must decide, which stimuli are relevant and whether and how to react to them. In the project MOBY-FLY we explore the question of how the brain initiates and appropriate behavioural response depending on sensory input using the fruit fly Drosophila as a model system. Specifically, we look at the control of rapid turns that flies perform during flight to avoid predators, collisions or change direction, which are called saccades. Saccades can be triggered by an approaching object (simulated in experiments using rapidly expanding "looming" stimuli) or occur spontaneously. These manoeuvres are extremely fast: Flies can change direction within about 50 ms reaching turning velocities of over 1000 degrees per second. However, flies do not always respond to looming stimuli with a saccade, and the direction of their turn varies depending on the perceived threat. Despite their importance for survival, the neural mechanisms controlling saccades remain largely unknown.
Our main objective is to identify the specific neurons in the brain responsible for saccade control and understand how they process information to trigger these rapid turns.
Previous research has shed light on how looming stimuli—one of the strongest triggers for saccades—are processed by the visual system. In earlier work, we identified a type of descending neuron, which transmits information from the brain to the motor system in the ventral nerve cord. This neuron’s activity is correlated with saccade execution in a head-fixed preparation. Using the available connectome, which provides information on how the neurons of the brain are connected, we try to identify candidate neurons within the central brain involved in the control of saccades. We then record their activity using electrophysiological and optical techniques in a head-fixed preparation, in which we detect saccades as rapid changes in wing stroke amplitude. In addition, we use the genetic tool kit available in Drosophila to manipulate the activity of the identified neurons to test whether and how they contribute to the initiation of a saccade. To do this under naturalistic conditions, we are developing a free flight setup allowing us to track the flies’ behaviour at high resolution.
With this work, we want to gain an in-depth understanding on how an important natural behaviour of a fly is controlled by its nervous system with a focus on how behavioural variability is generated. We hope that this will provide general insights into how the nervous systems selects and initiates an appropriate behavioural action, which could e.g. help the design of autonomously flying vehicles in the future.
Our main objective is to identify the specific neurons in the brain responsible for saccade control and understand how they process information to trigger these rapid turns.
Previous research has shed light on how looming stimuli—one of the strongest triggers for saccades—are processed by the visual system. In earlier work, we identified a type of descending neuron, which transmits information from the brain to the motor system in the ventral nerve cord. This neuron’s activity is correlated with saccade execution in a head-fixed preparation. Using the available connectome, which provides information on how the neurons of the brain are connected, we try to identify candidate neurons within the central brain involved in the control of saccades. We then record their activity using electrophysiological and optical techniques in a head-fixed preparation, in which we detect saccades as rapid changes in wing stroke amplitude. In addition, we use the genetic tool kit available in Drosophila to manipulate the activity of the identified neurons to test whether and how they contribute to the initiation of a saccade. To do this under naturalistic conditions, we are developing a free flight setup allowing us to track the flies’ behaviour at high resolution.
With this work, we want to gain an in-depth understanding on how an important natural behaviour of a fly is controlled by its nervous system with a focus on how behavioural variability is generated. We hope that this will provide general insights into how the nervous systems selects and initiates an appropriate behavioural action, which could e.g. help the design of autonomously flying vehicles in the future.
We have identified a descending neuron, DNp03, which receives input from neurons of the visual system that are responsive to looming, and characterized its activity. As expected from its connectivity, this neuron responds to looming stimuli on the side ipsilateral to its dendritic tree. Interestingly though, this neuron is mostly active during flight and does not fire action potentials when the fly is not flying. In addition, we could show that the neuron’s activity is to some degree correlated with whether or not the fly reacts to the looming stimulus with a saccade and exhibits a more sustained activity if the fly does. Therefore, its activity is reflective of the behavioural decision to turn. In the future, we want to explore, how this sustained activity is generated.
When we artificially activated DNp03 during free flight, the flies consistently performed an evasive response to change their flight trajectory. DNp03 also directly connects to descending neurons associated with spontaneous saccades. This suggests that multiple descending neurons work together to produce saccades.
We are currently identifying additional descending neurons involved in saccade initiation. To determine whether different neurons serve specialized roles, we are enhancing our free-flight tracking system to record flight manoeuvres with higher spatial resolution. By applying machine learning techniques, we aim to classify different types of responses and analyse how specific neurons contribute to distinct aspects of turning behaviour.
When we artificially activated DNp03 during free flight, the flies consistently performed an evasive response to change their flight trajectory. DNp03 also directly connects to descending neurons associated with spontaneous saccades. This suggests that multiple descending neurons work together to produce saccades.
We are currently identifying additional descending neurons involved in saccade initiation. To determine whether different neurons serve specialized roles, we are enhancing our free-flight tracking system to record flight manoeuvres with higher spatial resolution. By applying machine learning techniques, we aim to classify different types of responses and analyse how specific neurons contribute to distinct aspects of turning behaviour.
Most studies on insect flight have been conducted in head-fixed setups, where flies are physically restrained. While some research has examined freely flying flies, it has mainly focused on describing behaviour rather than linking it to specific neural circuits.
Our approach goes beyond this by combining precise genetic manipulation with high-resolution tracking of free-flight behaviour. This allows us to selectively activate or silence specific neurons and observe their contribution to flight manoeuvres.
Using this method, we have shown that activating just a single type of descending neuron is enough to trigger an escape response in a freely flying fly. This demonstrates that individual neurons can have powerful effects on behaviour, even in naturalistic conditions.
To further refine our understanding, we are improving our free-flight tracking system to achieve higher spatial accuracy. This will allow us to analyse subtle differences in flight manoeuvres and determine whether different types of descending neurons contribute to distinct aspects of saccade execution.
Our approach goes beyond this by combining precise genetic manipulation with high-resolution tracking of free-flight behaviour. This allows us to selectively activate or silence specific neurons and observe their contribution to flight manoeuvres.
Using this method, we have shown that activating just a single type of descending neuron is enough to trigger an escape response in a freely flying fly. This demonstrates that individual neurons can have powerful effects on behaviour, even in naturalistic conditions.
To further refine our understanding, we are improving our free-flight tracking system to achieve higher spatial accuracy. This will allow us to analyse subtle differences in flight manoeuvres and determine whether different types of descending neurons contribute to distinct aspects of saccade execution.