When asked to imagine a particular object the mind seems to conjure the image instantly and you can observe it's many features with your mind's eye. Given that we all do this dozens of times a day it seems like a fairly boring and menial occurrence. However if you stop to think of the mental processes that underlay this phenomena you'll see how complex and important it is.
To first test the manner in which mental images are formed participants were shown an empty grid with a lower case letter beneath it. They were asked to imagine the corresponding uppercase letter in the grid as quickly as possible and state whether the letter would cover a particular block in the grid. As the letters became more complex (containing more segments) the response time increased. This finding led the researchers to believe that an image was not formed all at once, but rather as individual parts. (the letter F should take longer to recall because you must recall three parts as opposed to an L which contains only two parts). It also appears as if the parts were imaged in almost the same order in all cases. This is based off the fact that when the subjects were asked to draw individual letters they were imaged in the same sequence at least 75% of the time.
These findings were further supported by auxiliary evidence of the brain forming patterns. For example ------ is viewed as a straight line, not six dashes. Similarly XXX XXX is viewed as two groups not six X's as is XXXooo. Thus the brain is predisposed to organizing things as parts or "perceptual units." Given this it seems likely that the brain stores the letter F by its three individual parts rather than as a whole, and when the letter is recalled it is imaged a part at a time based off previously stored perceptual units.
Why is this the case? Why not simply remember something as a whole rather than by its parts? The answer comes via the limitations of the brain.
Previous research has shown that it is more difficult to hold onto a mental image while you are paying attention to actual visual stimulation. This would seem to say that some of the cortices involved in visual sensory input are implicated in mental imagery. To test why the things are organized by parts the researchers look at the processing of visual stimuli.
The two primary visual cortices are located in the inferior temporal lobe and the parietal lobe. Ablations in monkeys of the inferior temporal lobe causes the monkey difficulty in discriminating against patterns and shapes but have no difficulties in object location. The reverse is true in animals with ablations of the parietal lobe indicating that each lobe has a different functional relevance in visual processing. Thus the observation of "what" and "where" are processed separately.
This explains why we image things sequentially. The shape of a part is stored separately from it's location relative to other parts. For example an F is composed of vertical bar connected on top and in the middle to two horizontal bars. Given that 94% of the participants drew an F by drawing the vertical line first then top and bottom shows that there are parts prerequisite to other parts and thus must be imaged one part at a time.
This division of objects into parts has great importance. It is why we are able to recognize letters in different fonts. Rather than memorizing how an F looks in every form and being confused upon seeing a new font we merely have to know the parts of an F and how they relate to each other. This can be further extrapolated out away from the simplicity of letters. A person can assume many forms. We can stand, sit, curl in the fetal position and still we are able to recognize it as a person because we recognize the parts and how they are connected to one another, as opposed to knowing exactly how a person looks when they are curled.
This division relieves us of having to know much more specific information thus freeing up brain power so we can say, know how to write a blog.
Aspects of a Cognitive Neuroscience of Mental Imagery. Kosslyn et al. Science Journal.