Neural field equations model population dynamics of large-scale networks of neu-rons. Wave propagation in neural fields is often studied by constructing traveling wave solutions in the wave coordinate frame. Nonequilibrium dynamics are more challenging to study, due to the nonlinearity and nonlocality of neural fields, whose interactions are described by the kernel of an integral term. Here, we leverage interface methods to describe the threshold of wave initiation away from equilibrium. In particular, we focus on traveling front initiation in an excitatory neural field. In a neural field with a Heaviside firing rate, neural activity can be described by the dynamics of the interfaces, where the neural activity is at the firing threshold. This allows us to derive conditions for the portion of the neural field that must be activated for traveling fronts to be initiated in a purely excitatory neural field. Explicit equations are possible for a single active (superthreshold) region, and special cases of multiple disconnected active regions. The dynamic spreading speed of the excited region can also be approximated asymptotically. We also discuss extensions to the problem of finding the critical spatiotemporal input needed to initiate waves.