It is widely accepted that spatial reasoning plays a central role in artificial intelligence, for it has a wide variety of potential applications, e.g., in robotics, geographical information systems, medical analysis and diagnosis. As noticed by many authors, spatial and temporal reasoning have a close connection. In this paper we propose a new, semidecidable, modal logic for spatial reasoning through directional relations, which is able to express meaningful spatial statements. Spatial Propositional Neighborhood Logic can be polynomially reduced to a decidable temporal logic based on time intervals preserving, at least, valid formulas. Thanks to such a reduction, we are able to re-use a sound and complete tableaux method in order to reason with Spatial Propositional Neighborhood Logic; to the best of our knowledge, there are practically no previous attempts of devising automatic reasoning methods for spatial reasoning.