So what is the brain doing to make us see?
This is the $64,000 question, isn't it.
When we look at cortical processing we might get the idea that the brain is extracting features from the visual stimulus. The receptive fields of cells in V1 are finding bars, edges and corners with specific orientations and directions of motion.
Perhaps the visual system builds up the representation of objects by detecting features that make up the object's contours. Or perhaps the pattern of response of a number of higher order features indicates the presence of a particular object.
What happens at later stages in this scenario? Are patterns of features used to indicate objects? Are there neurons later on in the visual system that respond to specific objects? Early on in the history of electrophysiology of the visual system, when Lettvin (1959) suggested that the ganglion cells in the frog's retina are fly detectors, actually "movement-gated dark convex boundary detectors" in his words, he was rebuked, laughed off the conference stage, and had difficulty in getting his work published. Since then we find a plethora of studies looking at cortical processing as feature processing.
Studies have even revealed neurons in the temporal cortex that respond specifically to hands and faces. The figure above is such an example. Do we have grandmother cells and yellow Volkswagen cells? Do we have yellow cells and different types of car feature cells that indicate yellow Volkswagens when the appropriate combinations are triggered simultaneously?
This is one approach to answering the $64,000 question.
Because many different cortical regions have been identified anatomically and physiologically (by differences in structure, cell response, and identification of topographic representations of the visual field) there has been a tendency to attribute different regions with different functional aspects of vision. For example area V4 has been identified as an area that is very responsive to color so therefore it must be the place where our experience of color is processed. Similarly, area MT is very responsive to moving stimuli leading to its identification as the motion center of the brain.
As we have seen by looking at diagrams of connectivity between cortical areas, the processing of vision is quite complex and partitioning areas of the brain as "color only" or "motion only" is probably a bit too simplistic. However, damage to specific regions in the brain can lead to specific losses in visual perception. Such deficits are well known in the areas of speech and language. For example, Zihl, et al. (1983), reported a patient who had a stroke that damaged a particular region homologous to area MT (MT is in the monkey brain). This patient could no longer see motion although her vision was otherwise normal. This is known as motion agnosia. There are several reports in the literature of cerebral achromatopsia, in which damage to areas that may be homologous to V4, lead to a loss in color vision so that the world looks achromatic.
Here are a couple of diagrams that exemplify this idea of cortical modularity:
The "What" and "Where" Pathways
|Ungerleider and Mishkin (1982) identified two processing stream that extend from striate cortex in the occipital lobe to the parietal and temporal lobes. The pathway that extends to the temporal lobe is attributed to be important for identifying objects. So the temporal pathway has been called the "what" pathway. The parietal pathway is attributed to be crucial for locating objects. This has been christened the "where" pathway.|
In their studies, they trained monkeys to perform two tasks. In one task, called object discrimination, the monkeys are presented with an unfamiliar object. They are then presented with this object and a new object. The monkeys task is to pick the newer object showing that they recognize the original object. If they choose the correct object they are rewarded. Monkeys quickly learn to perform this one-trial learning task. In the second task, landmark discrimination, monkeys are placed in front of a table that has two covered foodwells. A cylinder is placed next to one of the foodwells. If the monkeys choose the foodwell next to the cylinder they are rewarded (by getting the food). Monkeys quickly learn to pick the foodwell next to the cylinder.
Now after training, these monkeys have either bilateral removal of area TE in the temporal lobe or bilateral removal of the posterior parietal cortex. Those monkeys who have temporal lobe lesions are no longer able to perform the object discrimination task. Monkeys with parietal lesions can no longer perform the landmark discrimination task. In each case the monkeys are still able to perform the other task.
In humans, damage to regions in the temporal lobe can lead to deficiencies such as prosopagnosia, which is the inability to recognize faces. Prosopagnosics cannot even recognize the faces of their close friends and relatives however they can identify them by the sound of their voices. Damage to areas in the parietal lobe can lead to a condition known as contralateral or unilateral hemineglect in which the patient is unaware of the goings-on in the visual field contralateral to the side of the brain that has sustained the damage. For example, these patients may only eat the food on one side of their plate or only dress one side of their body. It is not simply a matter of a scotoma on one half of their visual fields since they can do better than chance in forced-choice experiments measuring visual ability in the blind hemifield. These patients however, do not report any conscious perceptions in this hemifield.
Computational Analysis of Cortical Functioning
In my own thinking about brain function, I am more inclined to wonder about the brain's computational methods than the mapping between perceptual features and tentatively identified visual streams. I find it satisfying to learn that the magnocellular pathway contains the best representation of high temporal frequencies, but less satisfying to summarize the pathway as the motion pathway, since high temporal frequency information may also be used in many other types of performance tasks. The questions I find fundamental concerning computation are "how?" not "where?" How are essential signal-processing tasks, such as multiplication, addition, and signal synchronization, carried out by the cortical circuitry? What means are used to store temporary results, and what means are used to represent the final results of computations? What decision mechanisms are used to route information from one place to another?
-Wandell, pp. 190-191 in your textbook