Animal motor control arises from the intricate interplay between neural circuits, biomechanics, and sensory feedback. Deciphering how these components interact is essential for uncovering the fundamental principles of movement generation. This dissertation investigates motor control in Drosophila melanogaster through a data-driven, integrative modeling approach synthesizing behavioral data, biomechanical simulations, and computational neural control.

To establish a biomechanical foundation, we introduce NeuroMechFly, a neuromechanical model of the adult fruit fly, reconstructed from a CT scan. This model serves as a versatile testbed for evaluating kinematic and neural controllers within a physics-based simulation. Expanding upon this framework, we developed a Hill-type muscle model for the fly's foreleg, incorporating anatomical and physiological data to generate biologically plausible muscle actuation. Through inverse dynamics, we analyzed muscle synergies and their contributions to coordinated limb movements during locomotion. Additionally, we enhanced the biological fidelity of NeuroMechFly by integrating multi-modal sensory processing and developing a novel inverse kinematics approach. These advancements significantly narrow the gap between real and simulated experiments in terms of biomechanics and environmental interactions.

Beyond biomechanics, motor control is ultimately orchestrated by the nervous system. In the final part of this thesis, we shift focus to the neural mechanisms underlying antennal grooming, a behavior with exquisite motor coordination. By integrating animal experiments, 3D kinematic analysis, and connectomic investigations, we found that grooming body part coordination is primarily governed by a feedforward mechanism controlled by central neurons in the brain. Simulating the connectome further uncovered distinct network motifs responsible for coordinating multi-limb movements, potentially conserved across behaviors.

Together, the methods and models developed in this dissertation provide a physiologically and anatomically grounded framework for investigating motor control in Drosophila from multiple perspectives. Furthermore, this work sheds light on how neural networks in the brain enable the flexible co-recruitment of multiple body parts to execute coordinated motor actions.