Background: Anomalous retinal correspondence (ARC) is a neural adaptation to eye misalignment in which non-corresponding retinal points are linked in the visual cortex to provide binocular fusion. ARC within the striate cortex would require that horizontal neurons link right-eye and left-eye ocular dominance columns (ODCs) separated by a distance in the cortex proportional to the angle of strabismus. Two hypothetical mechanisms are possible: (1) The ODCs can be linked by axons of horizontal neurons that project monosynaptically from a right-eye to a left-eye ODC. The further apart the ODCs, the longer the axons; hence, axon length should be greater in subjects with strabismus than in healthy subjects (elongated axon, monosynaptic hypothesis). In this case, the clinical probability of developing ARC should be independent of the angle of strabismus, until an upper-limit angle of strabismus is reached equally to the maximal length of axons available to link nonadjacent ODCs, at which point an abrupt decline of ARC probability should be evident. (2) Alternatively, ODCs can be linked by a chain of horizontal neurons, the number of which increases as the distance among ODCs increases; axon length in subjects with strabismus would be expected to be the same as in healthy subjects (normal axon, polysynaptic hypothesis). In this case, the greater the angle of strabismus, the more horizontal neurons and synapses required for linkage, and the greater the probability of signal degradation. Thus, the clinical probability of developing ARC through a polysynaptic mechanism should be inversely proportional to the angle of strabismus. The purpose of this study was to test these 2 hypotheses neuroanatomically in primates and clinically in children.
Methods: For the neuroanatomic portion of the study, biotinylated dextran amine was injected into ODCs of area V1 to label individual neurons. The length of the horizontal axons from these neurons was then compared in strabismic and normal monkeys. In the clinical portion of the study, the medical records of 192 children with strabismus were reviewed retrospectively. The angle of strabismus (prism cover test) and the presence of ARC (Bagolini striated lenses, Worth/Polaroid 4-dot) were recorded. Plots of the presence of ARC as a function of the angle of strabismus were obtained.
Results: There was no significant difference in axon length between healthy (7. 02 +/- 0. 83 mm) and strabismic monkeys (6. 60 +/- 1. 07 mm) (P =.16). In children with strabismus, ARC decreased as the angle of strabismus increased (P <. 05). ARC was more prevalent in children who had primary or postsurgical deviations of </=4 degrees to 5 degrees (8-10 PD), corresponding to </=2 horizontal axon lengths in the foveal striate visual cortex.
Conclusions: The visual cortex adapts to strabismus by combining information from paired ODCs of opposite ocularity that, because of the eye misalignment, are nonadjacent and separated by abnormally long distances across the striate cortex. The cortex appears to achieve the linkage, not by elongating neuronal axons, but by using chains of neurons that have normal-length axons. The visual cortex is most successful stochastically at achieving this linkage (ie, developing ARC) when the gap that must be bridged is no greater than 4 degrees to 5 degrees (8-10 PD), or the retinotopic distance in the foveal visual field is spanned by 2 normal V1 neurons.