Abstract
Oculomotor physiologists usually record one eye and assume that the other rotates identically (or oppositely during vergence). This approach is reflected in the structure of virtually all oculomotor models, as they generate either pure conjugate or pure vergence eye movements. However, many eye movements acquire targets that depart from the horopter or midline, requiring asymmetric eye rotations. How are asymmetric eye movements controlled? According to Hering’s (1868) theories, a unitary vergence command is simply added to a unitary conjugate command to drive gaze shifts to any location, and most models operate under this assumption. Our recent results during monocular viewing show asynchronous rotation of an occluded eye making it unlikely that a unitary vergence command drives gaze shifts to objects in depth (Chandna et al., 2021). Theoretical independent control of each eye could substitute for vergence and solve the problem, but pure independent control does not account for sluggish midline vergence saccades. Nor does it account for inappropriate conjugate rotations during gaze shifts to targets aligned with one eye (Müller paradigm). Here we introduce a new model of binocular control. At the core of the model is a rapid conjugate system that issues a unitary command to both eyes, and a pair of independent controllers which modulate that command’s effect on each eye. Like other oculomotor models, ours uses feedback to generate eye rotations with realistic temporal dynamics. The model generates an array of common and mysterious oculomotor behaviors including rapid saccades, asymmetric eye movements, slow midline vergence and inappropriate conjugate rotations in the Müller paradigm. A novel endogenous component interprets target motion when sensory evidence is lacking to explain our asynchronous covered-eye results.