Latent Binocular Interactions in Cortical Area V1 of Human Amblyopia

Binocular interaction in amblyopia is typically reported as being absent or greatly reduced. However, Baker et al, (2007) showed intact binocular summation by increasing stimulus contrast to the amblyopic eye in strabismic amblyopia. Here we used fMRI-informed EEG source imaging to measure latent binocular capacity in primary visual cortex (V1) by varying contrast in the amblyopic eye while keeping contrast fixed in the fellow eye. We made measurements in both anisometropic and strabismic amblyopes and compared their data to age-matched normal vision controls. We measured high-density steady-state VEP responses to two parallel grating stimuli (11° diameter, 2 cpd, sine-reversal) with different temporal frequencies presented dichoptically to the two eyes. The contrast in the amblyopic eye (F1=6.07Hz) was either swept from 1% to 40%, or fixed for a block of trials (at 40%, 20%, 10% or 0%, respectively), while the contrast of the fellow eye (F2=8.25Hz) was always fixed at 20%. The presence of intermodulation (IM) frequencies, which are sums and differences of the frequencies presented to the two eyes, was taken as evidence of binocular interaction. In the fixed-contrast condition, both ansiometropic and strabismic amblyopes showed the largest IM responses in V1 when the contrast in the amblyopic eye was 40% and contrast in the fellow was 20%. In the swept-contrast condition, the IM contrast-response function in V1 of anisometropic amblyopes was shifted to higher contrast levels, compared to normal controls. Strabismic amblyopes had significantly reduced IM responses, and only individuals with residual binocular fusion showed a weak IM response at high contrast in the amblyopic eye. Consistent with Baker et al. (2007), we demonstrate evidence for binocular interactions in early visual cortex in some, but not all, types of amblyopia. This finding suggests that amblyopes with residual fusion have the potential to recover deficits in binocularity.

Meeting abstract presented at VSS 2017