On the "cyclopean eye": saccadic asymmetry and the reliability of perceived straight-ahead

If two targets are both on the visual axis of one eye or the other, and binocular fixation is shifted from the farther one to the nearer, the aligned eye consistently makes an initial, seemingly pointless saccade in a temporal direction. The size of those saccades typically differs markedly, depending on whether the targets are aligned with the observer's dominant or non-dominant eye. Pickwell [(1972) Vision Research, 12, 1499-1507] proposed that this binocular asymmetry in oculomotor performance reflects a subject-specific lateral displacement of the egocenter (the "binoculus" of Hering, which has traditionally been assumed to be on the midline). An empirical test of Pickwell's widely endorsed hypothesis has now been conducted and the proposal has been found wanting. In an otherwise darkened room, subjects were required repeatedly to set a small light to a perceived straight-ahead location in the horizontal plane, first for a target at 300 cm distance and then for one at 30 cm. Extrapolation of a line that connects the two averages of those settings to the inter-ocular axis provides an estimate of the subjective egocenter to which visual directions are referred. Contrary to Pickwell's proposal, those locations of the inferred egocenter were usually quite near the midline, and were completely uncorrelated with same-subject data on the extent of saccadic asymmetry at the onset of asymmetrical convergence. The data on perceived straight-ahead underlying this result indicate the availability of extraretinal information about eye orientation that is quite precise at a given moment (median standard deviation of 47 min arc) but conspicuously non-stationary over several-minute intervals (monotonic drifts in sequential settings being very common).