Model-based study of the human cupular time constant

Abstract

The time constant of the angular vestibulo-ocular reflex (aVOR), measured from the response to steps of rotation about a yaw axis, has frequently been estimated as a single exponential. However, the slow phase velocity envelope during per- or post-rotatory nystagmus is more accurately represented by two exponential modes. One represents activity in the vestibular nerve induced by deflection of the cupula, the other by activation that the input from the canals produces in the central velocity storage integrator. The sum of the cupula and the integrator responses describes the overall response of slow phase eye velocity and can be approximated by a double exponential. Frequently, there is a plateau in the initial portion of eye velocity response, but this may be masked by habituation, making the cupula contribution unobservable and impossible to estimate. Using a model-based technique to analyze responses with a clear plateau, we estimated peripheral and central vestibular time constants by double exponential fits to slow phase eye velocity. Cupular time constants were varied from 1 to 10 s to identify values that gave optimal fits of the data according to a Chi-square criterion. The mean cupular time constant for 10 human subjects was 4.2 +/- 0.6 s. Fits of the data were also good for time constants between 3.5 to 7 s, but not for 1 to 3 or 7.5 to 10 s. The estimated cupular time constants also fit responses where there was no plateau. In 8 monkeys, cupular time constants were estimated as 3.9 +/- 0.5 s, which agreed with those derived from activity in the vestibular nerve. There was no difference between monkey and human cupular time constants from these estimates. It is likely that the human cupular time constant is similar to that of the monkey and shorter than previously thought.