A function called "spatial attention" is really important. The reason we have a unified perception of an object instead of seeing a bunch of disjointed color, shape, and other features is because we bind them together by directing "spatial attention" to a particular location. When spatial attention is compromised after damage to the parietal lobe, people can develop Balint's syndrome, an inability to notice more than one object at a time. This can make it difficult or impossible for them to take care of themselves or engage with other people or changes in their environment. Other types of parietal lobe damage can lead patients to neglect one whole side of space. It's not that half their visual field has been cut off--moving their heads doesn't do anything, and they neglect one side of whole objects, too. Rather, it's as if their spatial attention finds it difficult or impossible to move to that side of space. And on a more mundane level, poor performance on laboratory spatial attention tasks has been associated with ADHD and is probably also involved in Nonverbal Learning Disability.
But "spatial attention" is an odd hybrid concept of "attention" and "spatial perception," each of which have their own distinctive properties and are measured and understood very differently. Thus, it's hard to understand what spatial attention really means, and why it matters so much.
Attention is more like working memory: it's a limitation on the amount of information that can be held and worked with at a time. Attention, in some theories, IS the set of all information in conscious awareness. Limitations on attention, then, are quantitative limitations. The modality (verbal vs. visual) of attention may affect the nature of these quantitative limitations (e.g. a different limit might replace the famous, verbally-derived 7 + or - 2), but in essence, attention is best defined by the quantitative limits it puts on cognition.
Spatial perception, on the other hand, is a set of several maps or grids that we build by input from several different senses (hearing, sight, balance, proprioception), and multiple sources of information within the same sense (e.g. to compute depth from a 2D retinal image, we use some combination of stereoscopic, structure from motion, shading, texture, shadow, size, and perspective cues). We know there are several grids because space can be represented with reference to one's own body (egocentrically), with reference to particular objects, or with reference to cardinal directions (e.g., N/S/E/W). A problem with spatial perception seems like a difficulty representing space in one of these ways, or integrating these processes together to form a coherent mental map. These are fundamentally qualitative problems with cognition.
The question I'm asking is this: when spatial attention goes wrong, is the root cause a problem with attention or a problem with spatial representations? Or both? If both, do attention problems and spatial representation problems lead to different difficulties with spatial attention? And what kind of experiments can one use to answer these questions?