After having reviewed the vision system in detail, some aspects seem more clear. The forming of a visual concept involves the integration of smaller, low-level parts of the visual system. The smallest components, the photoreceptors, compose the ON- and OFF-center receptive fields of bipolar and ganglion cells. The ganglion cells then send action potentials onward through the LGN to the striate cortex, where first steps of integration happen in layers outside of layer IVC. Binocularity, meaning responsiveness to stimuli coming from either eye, happens in layers superficial to layer IVC. Another form of integration, happening outside layer IVC, is the integration of smaller receptive fields into bar-shaped fields. These fields can be selective of different features, e.g. orientation or direction of movement.
Progressing on to other cortical areas, they further integrate the information from lower structures, with different areas taking care of different operations. Area V5, receiving input from areas V2, V3 and direct axons from layer IVB of V1, is associated with motion selectivity. Areas running in the ventral stream, e.g. area 4, area IT and the fusiform face area are associated with color, shape and even face detection. Their real functioning is much more complex, as most areas receive signals from cells associated with most of these functions.
The idea behind concept forming is the gradually increasing integration of all of these parts of the system. In pattern recognition, many areas activate simultaneously, so there is no single area at the top of the hierarchy, responsible for coming up with the final image. With this said, my current view of concept forming resembles more the gradient-like manner in which colors are perceived: rather than a single set of neurons at a specific area firing on the recognition of a given pattern, maybe the image is formed more like a synchronized approximation involving a variety of areas.