Impulsive head rotation resets oculopalatal tremor: examination of a model

We have described a neuromimetic model of the interaction between the inferior olive (IO) and the cerebellum that accounts for symptomatic oculopalatal tremor (OPT), a disorder characterized by oscillations of the eyes (nystagmus), palate and other branchial muscles. OPT develops months after some brainstem strokes, in association with hypertrophic degeneration of the inferior olivary nucleus (IO). We hypothesized that OPT requires both (1) a pulsatile oscillator created by tighter electrotonic coupling between cells in the IO, and (2) a learned response from the cerebellar cortex that combines with the IO pulses to generate the quasi-pendular oscillations. Since the vestibular nuclei project to both IO and vestibulocerebellum, one prediction of the model is that rapid head rotations could interrupt the oscillator, effectively resetting the timing of the ocular nystagmus. The ocular oscillations in OPT vary in amplitude and phase, making it difficult to determine by Fourier analysis whether head perturbations phase-shift the nystagmus. We applied complex wavelet analysis to data from four patients with OPT and checked whether vestibular stimuli induced a change in phase of the nystagmus. First we calculated a threshold for the spontaneous rate of change of phase of OPT by comparing many segments of nystagmus waveform with their time-shifted versions, bootstrapping these arrays, and computing 95% prediction intervals for each patient. Then we compared the rate of change of phase due to each head perturbation with the threshold for that patient. To minimize the effects of the head perturbation itself on the wavelet analysis, we measured effects in a plane orthogonal to the head rotation, e.g., effects of horizontal head rotations on the torsional component of OPT. In all four patients, the rate of change of phase shift increased sharply at the time of the head perturbation, and in three the change was judged to be statistically significant. Thus, the experimental tests supported the prediction of our model for OPT.