Off-centric rotation axes in natural head movements: implications for vestibular reafference and kinematic redundancy

Until now, most studies concerning active head movements in three dimensions have used the classical rotation vector description. Although this description yields both the orientation of the head rotation axis and the amount of rotation, it is incomplete because it cannot specify the location of this rotation axis in space. The latter is of importance for a proper picture of the vestibular consequences of active head movements and has relevance for the problem of how the brain deals with the inherent kinematic redundancy of the multijoint head-neck system. With this in mind, we have extended the rotation vector description by applying the helical axes approach, which yields both the classical rotation vector as well as the location of the rotation axis in space. Subjects (n = 7), whose head movements were recorded optically, were instructed to shift gaze naturally to targets in 12 different directions at an eccentricity of 40 degrees. The results demonstrate that the axes for these head movements occupy consistently different spatial locations. For purely horizontal movements, the rotation axis is located near a point midway between the two ear canals. For gaze shifts in other directions, the rotation axes are located below the ear canals along two circles, one for movements with an upward component (up circle), the other (typically larger in size) for movements with a downward component (down circle). Purely vertical movement (up and down) axes were located on the lower pole of the up and down circles, respectively. It was found that both circles, the upper poles of which coincided, became larger in size as movement amplitude increased, which means that the axis location shifts to lower and more eccentric locations with respect to the skull for larger flexion and extension movements. Although this pattern could be recognized in most subjects, there were consistent intersubject differences in the absolute size of the circles, their increase with movement amplitude, and in the relative sizes of the up and down circles. Because multiple vertebrae are involved in head movements, there are theoretically many possibilities to execute a certain head movement. The differences in circle patterns among subjects indicate different strategies in resolving this kinematic redundancy problem, a fact that was not apparent from the classical rotation vector part of our description, which yielded a rather uniform picture. A simple model suggests that the downward shift of the location of the rotation axis requires a modulation in vestibulo-ocular reflex gain of