Deformation in Earth’s crust localizes onto faults that may rupture rapidly producing earthquakes or undergo slow aseismic slip. The detailed mechanisms that control the transition between the seismic and aseismic regimes and the onset of earthquakes remain unknown. These mechanisms control the geophysical processes preceding catastrophic failure, such as fracture development and strain localization on faults and in the rock volumes surrounding them. Our goal is to provide the first quantitative laboratory observations of the full displacement field in rocks before and during fault slip, and separate the aseismic and seismic components of it. We will develop novel experimental techniques based primarily on simultaneous dynamic synchrotron X-ray microtomography imaging and acoustic emission data acquisition and analysis. The data will reveal how slow and fast deformations develop and interact with each other in dry and wet crustal rocks under the stress, fluid pressure and temperature conditions at depths up to 10 km, and characterize the production of fractures during earthquake nucleation and rupture propagation. We will search for weak signals before dynamic rupture and develop ways to predict the time to failure from these signals. If we can demonstrate that the joint analysis of acoustic emission signals and X-ray microtomography data can be used to predict dynamic rupture in our experiments, we will have discovered an important lead towards earthquake prediction, which we will pursue in follow-up projects. We will compare the deformation microstructures produced in laboratory experiments with those of natural rock samples collected in California and Norway, where earthquakes occurred. The overarching goal is to progress toward a general model of the path to brittle failure in rocks by advancing knowledge of how fractures accumulate before and during both slow and fast earthquakes, under dry conditions and in the presence of water.