The Visual Pathways
The Visual Hemifields
The visual field, which is the total amount that can be seen with both stationary eyes, can be split into two halves, a right hemifield and a left hemifield. The central region of the visual field is the region that is viewed by both retinas, and is called the binocular visual field. Objects in the right hemifield fall on the left of both retinas, this being the nasal right retina and the temporal left retina. Nasal projections of both eyes decussate at the optic chiasm, the region where the 2 optic nerves converge. Therefore all information about the right hemifield projects to the left side of the brain.
The Retinofugal Pathway
The optic nerve contains the RGC axons, the sole output from the eye. It is the beginning of the retinofugal pathway, which projects away from the retina to the brainstem, primarily the lateral geniculate nucleus (LGN) of the thalamus – the gateway into the cortex. 90% of the fibres in this projection go to the LGN. The remaining 10% go to the pretectum for the photopupillary reflex (retinotectal projection), the hypothalamus for circadium rhythm, or the superior colliculus for eye movement.
The optic nerves converge at the optic chiasm, and this is the site of a partial decussation of the RGC axons. Axons of nasal RGCs cross to the contralateral side of the brain. Therefore nasal RGC fibres of the left eye combine with temporal RGC fibres of the right eye, and vice versa. These then travel in the optic tract to the ipsilateral LGN. Here they synapse on thalamic neurons that travel in the internal capsule as the optic radiation. These finally end mainly in the primary visual cortex (V1) of the occipital lobe. This region is largely located in the medial surface of both cerebral hemispheres, in the calcarine sulcus, and so is relatively hidden from view. From here, fibres project to the higher visual areas, such as V2 and V3 (brodmanns area 18) and V4 and V5 (brodmanns area 19).
For an image of the retinofugal pathway, click here.
At the LGN, the different kinds of inputs are kept separate. The LGN contains 6 layers (1-6), 4 of which are targets of the p-type RGC cells (3-6), and 2 of which are targets of the m-type RGCs (1 and 2). 1 is most ventral, 6 is most dorsal. In the right LGN, right temporal RGC axons synapse on LGN cells in layers 2, 3 and 5. Left nasal RGC axons synapse on the right LGN cells in layers 1, 4 and 6. So, right and left retinal inputs are kept separate, as are magnocellular and parvocellular inputs. Areas in between each layer, called koniocellular layers, receive non-m/p RGCs. Therefore different areas of the LGN are concerned with different visual information. The properties of magnocellular and parvocellular LGN neurons are similar to those of the corresponding RGCs. ‘On’ and ‘off’ cells are intermixed, but left/right and m-type/p-type information is kept separate. These inputs then project to V1 in the internal capsule, which is a white matter tract.
The LGN. This image is taken from http://commons.wikimedia.org/wiki/File:Lateral_geniculate_nucleus.png under the GNU Free Documentation License
The Primary Visual Cortex
Also known as the striate cortex (due to the defined stripe in layer 4 of the cortex), brodmanns area 17, or V1, the primary visual cortex contains 6 layers of cell bodies in grey matter that the axons of the optic radiation synapse onto. Approximately 1.5 million retinal neurons diverge onto approximately 200 million cortical neurons in V1, and a retinotopic map is maintained throughout this pathway.
Of the 6 layers, layer 4 is the target of the underlying white matter projections. Layer 4 is further divided into 4A, 4B and 4C. 4C is subdivided into 4Cα (magnocellular input) and 4Cβ (parvocellular input) and contains target stellate cells. Pyramidal cells in layers 2, 3 and 4B project to higher visual areas, whilst layer 6 pyramidal cells project back to the thalamus. The higher visual areas are also called the prestriate cortex and the visual association cortex. Pyramidal cells in layer 5 project to the superior colliculus in the midbrain tectum to control eye movement.
Left and right LGN inputs are separated when they reach layer 4C of V1 (they are monocular), and are split into equally spaced patches called ocular dominance columns, which lie parallel to the cortical surface and alternate left, right, left... Radial and horizontal connections are made between the different layers, and information in the retinotopic map is maintained such that a cell in layer 4C receives information from the same part of the retina layer 3 above it.
Layer 4Cα pyramidal cells (magnocellular) project axons upwards to layers 4B and layer 4Cβ pyramidal cells (parvocellular) project upwards to layer 3, where left and right inputs mix for the first time. Cells now have 2 receptive fields and are called binocular. Layer 4Cα cells are insensitive to colour, whereas layer 4Cβ are sensitive to colour. These pyramidal cells project out of V1 as previously stated.
There are 2 classes of cells in V1, simple cells and complex cells. These different classes of cells have different properties, and respond differently to orientation of edges, moving objects, and length of objects. Most are complex, and reside in layers 2, 3 and 5.
As well as the ocular dominance columns running parallel to the cortical surface, there are also orientation columns running perpendicular to the cortical surface containing simple cells, which are sensitive to the orientation/shape of objects.
There are also blob receptive fields that receive input from the koniocellular cells. These are dotted through ocular dominance columns and are sensitive to colour (wavelength sensitivity), but not to orientation or direction of objects. They are outside of layer 4. Interblob regions are in between the blob regions, and have many properties, such as binocularity and directional selectivity, but not wavelength selectivity.
The Visual Association Cortex
The visual association cortex contains V2, V3, V4 and V5 areas. It can be further simplified into 2 pathways, called the dorsal and the ventral pathways. The 2 clearest target areas are V5 and V4.
The Dorsal Pathway
The dorsal pathway is the ‘where’ pathway, and is dominated by magnocellular input (fast responding motion detecting cells with large receptive fields that are not sensitive to the wavelength of light – colour blind). This dorsal pathway is also involved in depth perception, which is called stereopsis. The pathway is from V1 to V2 to V5. V5, the final target of the dorsal pathway is the medial temporal lobe. From here information may proceed to the posterior parietal cortex. Other targets from V5 are the superior colliculus for eye movement.
V2 is divided into dark thick and dark thin stripes, with paler interstripes in between. The dark thick stripes of V2 are the part of V2 involved in the dorsal pathway. Therefore the dorsal pathway is as follows.
- V1 layer 4Cα – layer 4B – thick stripe of V2 – V5/medial temporal lobe.
The Ventral Pathway
The ventral pathway is the ‘what’ pathway, and is dominated by parvocellular input of which there are 2 parallel streams. The first is the parvo-blob stream. This is from V1 to the thin stripe of V2 to V4, and is involved in colour perception. The second is the parvo-magno stream. This is from V1 to the interstripe regions of V2 to V4. The final destination is the inferotemporal cortex. The parvo-magno stream is concerned with the perception of edges and fine form.
Therefore the ventral pathways are as follows.
- Parvo-blob stream. V1 layer 4Cβ – blob – thin stripe of V2 – V4.
- Parvo-magno stream. V1 layer 4Cβ – interblob – interstripe of V2 – V4 – inferotemporal cortex.