Plasticity and stability of visual field maps in adult primary visual cortex

Abstract

It is important to understand the balance between cortical plasticity and stability in various systems and across spatial scales in the adult brain. Here we review studies of adult plasticity in primary visual cortex (V1), which has a key role in distributing visual information. There are claims of plasticity at multiple spatial scales in adult V1, but a number of inconsistencies in the supporting data raise questions about the extent and nature of such plasticity. Our understanding of the extent of plasticity in V1 is further limited by a lack of quantitative models to guide the interpretation of the data. These problems limit efforts to translate research findings about adult cortical plasticity into significant clinical, educational and policy applications.

Figure 1. The ocular dominance columns and visual field map in primary visual cortex (V1)

(A) A medial view of the posterior right hemisphere of a post-mortem human brain. Human V1 is located principally in the calcarine sulcus, though its full extent frequently goes to occipital pole onto the ventral-lateral surface. CS - calcarine sulcus; CC corpus callosum. (B) The white/gray matter surface measured using MRI in a living subject. The surface rendering is inflated to increase the visibility of the sulci; it is shaded to emphasize the sulcal (dark) and gyral (light) regions. (C) A flattened post-mortem brain showing right calcarine and surrounding cortex from a subject with an enucleated left eye. The outlined region is V1. The cytochome oxidase staining forms light and dark bands that reveal the ocular dominance columns. The dark spot (arrow) is the projection zone from the left eye's blind spot (optic disk). (D) Calcarine and surrounding cortex computationally flattened from the MR-derived surface mesh. The color overlays identify the stimulus angle (left) or eccentricity (right) that most effectively stimulates each cortical location (measured using fMRI). Angle and eccentricity (up to 12 deg from the fovea) are measured with respect to fixation. The angle and eccentricity maps together define the V1 visual field map. The boundary between V1 and V2 can be identified in the angle map from the locations that respond best to the vertical meridians (arrows). Scale bars: 1 cm. [[I think it might be better to put 3 arrowhears instead of one outlining the borders between V1/V2 in this figure (i.e the head of the arrows would trace the border). As it is for the unfamiliar reader it may be difficult to figure out what the arrow points to]]

Figure 2. V1 neurons receive a diverse source of inputs

A V1 neuron may receive input from the lateral geniculate nucleus (LGN), extrastriate cortex (V2, V3, MT and other extrastriate sources), lateral connections between V1 neurons, and the pulvinar a large nucleus in the thalamus. In healthy V1, the reported receptive field (RF) size can differ by a factor of 4 depending on the nature of the mapping stimulus. (A) Inputs to V1 neurons have a wide range of receptive field sizes. The receptive field (RF) size of center-surround LGN inputs (bottom) is small compared to the RF size of extrastriate sources (top). Extrastriate sources have receptive field sizes that vary and can be larger than 5 deg in diameter . V1 neurons can also receive input from other V1 neurons with RF centers separated by a degree or more. The variations in estimated RF size probably result from different contributions from the pathways that deliver the input signals to the V1 neuron. (B) From the V1 visual field map, it is possible to express estimates of the RF center radius on the cortical surface. The radius of V1 RFs is often larger than 3mm; more than 10% of the neurons have a radius exceeding 5mm (From REF124, Fig 13A). The surround influence generally extends beyond 7mm.

Figure 3. No reorganization of the V1 visual field map after eye rotation

(A) Normally locations in the two eyes that receive corresponding images send their signals to corresponding locations in cortex. (B) Following surgical rotation of one eye, V1 inputs from the two eyes no longer represent corresponding visual field locations (solid lines). In principle, developmental plasticity could compensate for this misalignment by reconnecting retina and cortex (dashed line). Developmental plasticity does not correct the inappropriate mapping caused by eye rotation; the original, and now incorrect, mapping is preserved (solid lines).

Figure 4. V1 responses in humans with central retinal lesions

The four main images show an expanded posterior view of the calcarine sulcus (box on the upper inset). The color overlays compare fMRI responses in a juvenile macular degeneration (JMD) subject with a large central scotoma and spared vision in the lower peripheral field (left) and a Control with a similar “artificial” scotoma (right). In the Passive condition subjects passively viewed a visual stimulus presented in the peripheral visual field near the lower vertical meridian. In both subjects this produced a modest response in anterior calcarine at the location corresponding to the position of the stimulus in the peripheral visual field near the lower vertical meridian (upper images; arrows). In the Active condition subjects were asked to remember the visual stimulus from trial-to-trial (lower images). In this condition responses in the JMD subject spread significantly towards the occipital pole, and responses increased in other regions, such as the ventral surface. But in the Control there was no significant expansion of the BOLD signal into posterior calcarine. In both the JMD and the Control the Active task increased responses broadly, including near the occipital pole (arrow). The location of this activation with respect to V1 has not yet been defined with any certainty. The color bar indicates the amplitude of the BOLD response (percent modulation), either in synchrony (red) or out of synchrony (blue) with the stimulus. Only modulations exceeding 0.3% coherence are shown.

Figure 5. The expected effect of retinal lesions on V1 responses

(A) This schematic illustrates the diverse receptive fields of neurons expected to be found within a region of V1. The black circles show the size of neurons' RFs plotted on a representation of the visual field. The RF sizes vary and (partly) overlap. (B) The same RFs are shown with a transparent gray rectangle that indicates the portion of the visual field blinded by a simulated retinal lesion. The retinal lesion is located in the center of the RFs that are sampled from this part of cortex. The effect of the retinal lesion is to reduce the number of responsive neurons within the LPZ. Assuming no cortical plasticity, we still expect some cells to continue to respond to signals placed on adjacent regions of spared retina (red circles). Such neurons will necessarily have RFs that are displaced (ectopic) from their pre-lesion position. A reduction in the number of post-lesion responsive cells and ectopic RFs of these cells have been observed. Such data should be construed as supporting adult cortical plasticity only if the reduction in number of responsive cells and the change in the properties of the ectopic RFs differ significantly from a model that assumes no plasticity. A complete model should include quantitative specification of the RF size distribution, experimental factors (retinal penumbra due to inflammation or swelling), and models of retinal plasticity.