Maintaining spatial body alignment on a rotating platform by means of active counter-circling: role of vestibular and podokinesthetic afferents.

We investigated the behaviour of vision-deprived human subjects who try to maintain their horizontal alignment in space on a rotating platform by stepping about their own axis in counter-direction ('podomotor counter-rotation'), and we ask which of two alternative hypotheses best explains this behaviour. (1) The feedback hypothesis assumes that the podomotor counter-rotation is controlled by negative feedback of vestibular signals from the horizontal canals. (2) The reconstruction hypothesis holds that the vestibular cue first is combined with leg proprioceptive afferents signalling the individual's rotation on the platform ('podokinesthetic cue') in a way that reconstructs the platform's motion in space for internal representation; a negative (direction-inverted) copy of this representation then would drive the counter-rotation. Subjects were exposed to three different velocity profiles of platform rotation: VC, constant velocity rotation with sudden onset and offset; VS, sinusoidal rotation; VN, pseudorandom noise sequences. The subjects' response (i.show $132#e., their active self-rotation on the platform) to the onset and offset of VC rotations was reminiscent of a first-order lead system. Specifically, after rotation onset subjects immediately began to step on the platform in opposite direction; initially, the velocity of this response matched that of platform rotation, leading to a fairly good stabilisation of subjects' alignment in space. However, this response declined exponentially; consequently, subjects began to increasingly rotate in space along with the platform, ultimately stepping in place on the platform. After rotation offset, subjects immediately began to step around on the now stationary platform so as to continue their previous rotation in space; this response again declined exponentially until subjects became gradually stable again with respect to space. Within subjects, the time constant (tau) of these responses was similar for onset and offset. Across subjects it exhibited a conspicuous variability, ranging from 7 s to virtually infinity. The responses to VS and VN rotations were closely correlated to what could be predicted for each individual from his tau during VC on the assumption of a first-order lead system. We conclude that the mechanism stabilising body orientation basically is linear (no prediction with sinusoidal rotation, no extrapolation of constant velocity rotation). A comparison of the experimental results with simulations of the feedback hypothesis and of the reconstruction hypothesis suggests that the reconstruction hypothesis is a more likely description of the underlying processing of the vestibular and podokinesthetic cues.