The fast phases of vestibular nystagmus do not simply return the eye to the centre of the orbit, but also direct the eye beyond this primary position, in the anticompensatory direction. This induces a shift of the ocular beating field (Schlagfeld). This shift of the beating field has been mainly documented in studies of eye-head co-ordination in monkeys and cats. From studies in the cat, the functional role of this shift was interpreted as an anticipatory orienting response towards a "centre of interest". We investigated in humans whether this shift could be related to some strategy of orientation.

Eye movements were measured with an infrared system during two experiments on self-motion perception in the dark. Subjects were asked to rotate, by means of a joystick, a mobile robot on which they were seated in order to reproduce a previously imposed passive rotation (80 - 340°). In the first experiment, no specific instruction concerning orientation strategies was given. From this experiment, we made the hypothesis that the shift of the ocular beating field could be the manifestation of two different orientation strategies based on allocentric and egocentric reference frames, respectively. It was also proposed that subjects who preferably used the first strategy exhibited large shifts of the beating field, while the others who probably used egocentric memory did not exhibit any shift.

To further test this hypothesis in a second experiment, we gave to the subjects consecutively two different instructions: in the first set of runs, naive subjects, who had no mental image of the experimental room since they had never seen it before, were asked to pay attention exclusively to their sensations of body rotation in place, during both stimulus and reproduction rotations. Then, subjects were shown the room, and during the second set, subjects were asked to try to 'see' themselves turning in the room (to visualise mentally the room). Both during the stimulus and the reproduction rotations, mean gaze shifts were indeed larger in the second condition (5.73 ±17.03 deg; 6.32 ±17.05 deg) than in the first one (1.05 ±15.63 deg; 1.80 ±17.44 deg).

Therefore, we believe that the shift of the beating field could be influenced or could even participate in an orientation strategy when this strategy is based on an internal representation of the experimental room.