Inertial representation of angular motion in the vestibular system of rhesus monkeys. II. Otolith-controlled transformation that depends on an intact cerebellar nodulus

1. We recently studied the spatial representation of angular motion signals in rhesus monkeys by examining the orientation of postrotatory vestibuloocular responses during tilt of the head and body relative to gravity after constant-velocity rotation about an earth-vertical axis. We have reported that low-frequency angular motion signals in the vestibuloocular reflex (VOR) of rhesus monkeys are spatially transformed such that they remain invariant relative to gravity. In the present study we examine the properties of these inertial vestibular signals by employing similar stimulation conditions in animals with either selective semicircular canal plugging or selective lesions of cerebellar lobule X (nodulus) and ventral lobule IX (uvula). 2. We studied the spatial organization of postrotatory VOR in two rhesus monkeys that had either the lateral or one of the vertical canal pairs inactivated by plugging. In both monkeys, the spatiotemporal characteristics of postrotatory velocity after rotation in the plane of an intact canal pair and tilting in the plane of the plugged canal pair were indistinguishable from those of intact animals: postrotatory responses after tilts in the plane of the plugged canal pair were strongly damped, whereas an orthogonal response component was generated that rotated the eye velocity vector toward alignment with gravity. Thus otolith information rather than transient semicircular canal inputs that normally coexist during tilts seem to provide the necessary cues for the central transformation of semicircular canal signals. 3. We studied the three- dimensional VOR properties in two animals in which the cerebellar nodulus and ventral uvula were surgically ablated. After these lesions the temporal properties of the horizontal, vertical, and torsional VOR during earth- vertical-axis rotations were differentially affected. For horizontal VOR, the duration of postrotatory nystagmus was prolonged and the responses acquired strongly underdamped (i.e., oscillatory) properties. Similarly, sinusoidal responses were characterized by smaller phase leads after the lesion. For torsional VOR, the duration of postrotatory nystagmus was significantly shorter after the lesions, reaching postlesion values of 3.6 ± 1.7 (SD) s and 6.4 ± 1.1 s compared with prelesion values of 22.4 ± 4.5 and 33.6 ± 5.3 s for each animal. In addition, large phase leads characterized the torsional VOR during low frequency sinusoidal stimulation. The dynamic properties of the vertical VOR in the lesioned animals, on the other hand, were indistinguishable from those in controls. 4. The cerebellar lesions affected the spatial organization of the horizontal and vertical/torsional systems in a differential way. Inertial transformation of lateral canal activity was only partially affected. Head tilts during postrotatory horizontal VOR still induced generation of an orthogonal component in the appropriate direction; however, appropriate temporal tuning of the responses was lost. As a result, postrotatory eye velocity rotated toward alignment with gravity but systematically undershot the appropriate tilt angle. 5. Lesions of the nodulus and ventral uvula eliminated the ability of the vertical canal system to modify its dynamic characteristics as a function of head tilt. In contrast to the lateral canal system, however, this loss was also accompanied by a complete inability to reorient eye velocity relative to gravity. We conclude that neural processing of vestibular signals in the nodulus/ventral uvula is involved in maintaining an inertial coding of angular motion in the vertical canal system. 6. These experimental observations are simulated by a simple model in which semicircular canal activity is processed through an inertial center that reorients semicircular canal head velocity signals relative to gravity on the basis of a combination of rotation and projection principles. One of the predictions of the inertial vestibular system is the previously unexplained generation of a steady-state compensatory nystagmus during constant-velocity, earth-vertical-axis rotation when the head is simultaneously sinusoidally oscillated about a nested orthogonal axis (e.g., pitch while rotating). In agreement with this, animals with lesions of the nodulus/ventral uvula were unable to generate this response. 7. We conclude that there exists a central vestibulomotor system that utilizes both otolith and semicircular canal signals to detect absolute angular motion of the head in space. These results suggest that the vestibuloocular velocity storage merely represents an oculomotor footprint of this system whose function should be further addressed in head-free and freely moving subjects.