Spatial orientation of the vestibular system: dependence of optokinetic after-nystagmus on gravity

1. Monkeys received optokinetic stimulation at 60 degrees/s about their yaw (animal vertical) and pitch (animal horizontal) axes, as well as about other head-centered axes in the coronal plane. The animals were upright or tilted in right-side-down positions with regard to gravity. The stimuli induced horizontal, vertical, and oblique optokinetic nystagmus (OKN). OKN was followed by optokinetic after-nystagmus (OKAN), which was recorded in darkness. 2. When monkeys were tilted, stimulation that generated horizontal or yaw axis eye velocity during OKN induced a vertical or pitch component of slow phase velocity during OKAN. This has been designated as "cross-coupling" of OKAN. Eigenvalues and eigenvectors associated with the system generating OKAN were found as a function of tilt. They were determined by use of the Levenberg-Marquardt algorithm to minimize the mean square error between the output of a model of OKAN and the data. 3. The eigenvector associated with yaw OKAN (yaw axis eigenvector) was maintained close to the spatial vertical regardless of the angle of tilt. The eigenvector associated with pitch OKAN (pitch axis eigenvector) was always aligned with the body axis. The data indicate that velocity storage can be modeled by a piecewise linear system, the structure of which is dependent on gravity and the yaw axis eigenvector, which tends to align with gravity. 4. Yaw axis eigenvectors were also determined by giving optokinetic stimulation about head-centered axes in the coronal plane with the animal in various angles of tilt. A technique using a spectral analysis of residuals was developed to estimate whether yaw and pitch OKAN slow phase velocities decayed concurrently at the same relative rate and over the same time course. The eigenvectors determined by this method were in agreement with those obtained by analyzing OKAN elicited by yaw OKN. 5. During yaw OKN with the animal in tilted positions, the mean vector of the ensuing nystagmus was closer to the body axis than to the spatial vertical. This suggests that there is suppression of the cross-coupled pitch component during OKN. The direction of the stimulus may be utilized to suppress components of velocity storage not coincident with the direction of stimulus motion. 6. There were similarities between the monkey eigenvectors and human perception of the spatial vertical, and the mean of eigenvectors for upward and downward eye velocities overlay human 1-g perceptual data.(ABSTRACT TRUNCATED AT 400 WORDS)