1: \begin{abstract}

2:   Neural field models are nonlinear integro-differential equations for the evolution

3:   of neuronal activity, and they are a prototypical large-scale, coarse-grained

4:   neuronal model in continuum cortices. Neural fields are often simulated

5:   heuristically and, in spite of their popularity in mathematical neuroscience,

6:   their numerical analysis is not yet fully established. We introduce generic

7:   projection methods for neural fields, and derive a-priori error bounds for these

8:   schemes. We extend an existing framework for

9:   stationary integral equations to the time-dependent case, which is relevant for

10:   neuroscience applications. We find that the convergence rate of a projection scheme for a

11:   neural field is determined to a great extent by the convergence rate of the

12:   projection operator. This abstract analysis, which unifies the treatment of

13:   collocation and Galerkin schemes, is carried out in operator form, without

14:   resorting to quadrature rules for the integral term, which are

15:   introduced only at a later stage, and whose choice is enslaved by the choice of the

16:   projector. Using an elementary timestepper as an example, we demonstrate that the

17:   error in a time stepper has two separate contributions: one from the projector, and

18:   one from the time discretisation. We give examples of concrete projection methods: two

19:   collocation schemes (piecewise-linear and spectral collocation) and two Galerkin

20:   schemes (finite elements and spectral Galerkin); for each of them we derive error

21:   bounds from the general theory, introduce several discrete variants, provide

22:   implementation details, and present reproducible convergence tests.

23: \end{abstract}

24: