Ocular torsion (OT) in response to static or dynamic head tilt is generally considered to reflect functioning of the otolith organs or semicircular canals, respectively. Cervical contributions are either ignored or avoided by examining the response during tilt of the whole body. However, for horizontal eye movements it has been shown that the neck proprioceptors are responsible for anti-compensatory gaze shifts. The aim of the present study was to investigate whether similar effects exist for the OT response.

To differentiate between inputs from the otoliths, the semicircular canals and the neck proprioceptors, we used three sinusoidal roll tilt stimuli (the head alone, the trunk alone and the whole body) both in an upright and supine body orientation. The stimulus frequency in these dynamic conditions was 0.2 Hz, and the amplitude was 25°. The OT response was recorded in five healthy subjects by means of video-oculography with a sample frequency of 50 Hz.

The results showed a clear cervical contribution which was best noticeable by comparing the raw recordings obtained in conditions with only head tilt and conditions with whole body tilt. Although there was no difference in the slow component velocity (SCV), the amplitude of the OT nystagmus was significantly smaller during head tilt. Rotation of the neck generated anti-compensatory saccades, reducing the torsional excursions of the eyes. This is analogous to the cervically induced gaze shift observed with horizontal eye movements. On average, the cumulative amplitude of this cervical component amounted to 2.5°. Interestingly, the OT response to rotation of the trunk alone was much smaller (mean amplitude 0.4°). In fact, trunk rotation only produced an OT response in the upright body orientation, but not in the supine body orientation. These findings suggest that at least some vestibular stimulation is required for the cervical effects to become visible.

In addition to the sinusoidal roll stimuli, we also studied the OT response to static tilt of only the head, only the trunk or the whole body. This yielded similar data, although of smaller amplitude, confirming that the influence of the neck proprioceptors depends on the amplitude of the vestibular response. We therefore conclude that the cervical contribution to OT is to modulate eye position - not velocity - during an existing vestibular response, so that the eyes remain oriented in a head-fixed frame of reference. The usefulness of this cervical control may be explained as follows. During normal locomotion, the vestibular response produces ocular counterrolling which sufficiently stabilises the visual surround. However, during head movements to intentionally look at an object, it would be more appropriate to display the object straight (i.e. not tilted) on the retina. In that case, the vestibular response should be neutralised, which is achieved by the anti-compensatory cervical response.