Gaze-Accuracy during Monocular and Binocular Viewing

Abstract

Epelboim et al (1996 Vision Research 35 3401 – 3422) reported that cyclopean gaze errors were smaller than either eye's during tapping and looking-only tasks. This raised two questions: (i) does cyclopean gaze accuracy require binocular input, and (ii) when only one eye sees, is its gaze more accurate than the patched eye's? Most oculomotorists probably expect an affirmative answer to both. Neither expectation was fulfilled.

The Maryland Revolving Field Monitor recorded, with exceptional accuracy, eye movements of two unrestrained subjects tapping or only looking, in a specified order, at four randomly positioned LEDs, with monocular or binocular viewing. Subjects either tapped with their finger tips naturally, or unnaturally via a rod (2 mm diameter, 1.5 cm long), glued to a sewing thimble. Instructions were to be fast, but make no order errors.

With binocular viewing, cyclopean gaze accuracy was best during looking-only. During natural tapping, gaze errors increased, becoming no smaller than success required. Both tasks were learned equally fast, but as expected, the younger subject (aged 27 years) performed ∼ 40% faster than the older subject (aged 69 years). Unnatural, monocular viewing produced odd results, eg cyclopean gaze error was smallest when only one eye could see in some conditions. Only the older subject served in the unnatural tapping task because the younger's errors were too close to his gaze control limit. The older subject, who was suitable, reduced his cyclopean gaze error by 56%, from 1.4 to 0.9 deg. These results support our claim that the gaze error allowed is adjusted to the visuomotor demands of different tasks.