The vestibulo-ocular reflex: is it an independent subsystem?

Traditionally, the vestibulo-ocular reflex (VOR) is described as a distinct, phylogenetically old, oculomotor subsystem, which serves to stabilize gaze direction. It is supposed to act as a stereotyped reflex with definite input-output relations, which can be measured by rotating a subject passively in darkness, and which are kept at a nearly ideal level by adaptive parametric adjustments. This paper argues that such a view of the VOR may be not realistic: 1) the VOR in darkness (especially in humans) does not behave as a well-calibrated system; it has a low and variable gain which can be changed easily, even by purely mental assumptions; 2) a hard-wired VOR does not lead to appropriate eye movements in most natural situations, and would need continuous conditioning by other systems. As there is no compelling physiological or anatomical evidence for an independent VOR, it seems more fruitful to hypothesize that vestibular signals are just one of many inputs to a spatial localization process, which computes the relative motion between a subject and an object of his choice on the basis of all available, relevant information. Instead of a distinct subsystem, vestibulo-ocular responses in darkness may represent nothing more than the (somewhat arbitrary) default performance of this larger gaze-control system, functioning poorly in the absence of complete information.