Nematic liquid crystals, which are fluids with orientational order, may contain topological defects called disclinations where this order breaks down. While they’re undesirable in LCD screens, disclinations play an essential role in the dynamics of active nematics, non-equilibrium systems with internally driven flows coupled to nematic orientational distortions. Examples include cytoskeletal biofilaments with molecular motors, bacterial colonies, and some eukaryotic cellular tissues. In quasi-2D confinement, disclinations are point-like, and their pair-unbinding and motility drives the chaotic dynamics. In this talk, I will explore how the situation in 3D is even more complex. I will describe our efforts to model the first experiments on bulk 3D active nematics, in which we show that the dominant topological excitations are a certain family of topologically neutral disclination loops. We explore fundamentally 3D effects such as defect self-annihilation, recombinations between multiple defects, and twist distortions, as well as the consequences for 3D active flows.