David L. Tomko

Effect of gravity on vestibular neural development

The timing, molecular basis, and morphophysiological and behavioral consequences of the interaction between external environment and the internal genetic pool that shapes the nervous system over a lifetime remain important questions in basic neuroscientific research. Space station offers the opportunity to study this interaction over several life cycles in a variety of organisms. This short review considers past work in altered gravity, particularly on the vestibular system, as the basis for proposing future research on space station, and discusses the equipment necessary to achieve goals. It is stressed that, in keeping with the international investment being made in this research endeavor, both the questions asked and the technologies to be developed should be bold. Advantage must be taken of this unique research environment to expand the frontiers of neuroscience.

Adaptive plasticity in the naso-occipital linear vestibulo-ocular reflex.

Abstract The linear vestibulo-ocular reflex (LVOR) during motion along the naso-occipital (NO) axis is governed by eye position and viewing distance. These influences are necessary for the LVOR to maintain stable foveal images during head translation. The response to NO translation must be large when eye position is eccentric from the axis of head motion (i.e., during lateral gaze) and must diminish as eye position approaches straight-ahead, eventually reaching zero when the eye is aligned with the NO axis of motion (the ”null point”). As eye position crosses to the opposite side, the LVOR response must reappear, but in the opposite direction, and must grow in magnitude as eccentricity increases. To determine whether the NO-LVOR is subject to adaptive plastic mechanisms, squirrel monkeys were conditioned during NO translation while they binocularly viewed a rich visual field through parallel base-right or base-left wedge prisms. This optical method effectively shifted the visual world 9° leftward or rightward, respectively, thus inducing a mismatch between vision and the NO-LVOR during head movements. To restore compensatory function, the relationship between LVOR sensitivity and horizontal eye position must shift by 9° in the same direction as the visual image shift, effectively shifting the null point. After 2 h of adaptive conditioning, all monkeys exhibited an adaptive shift in the appropriate direction by an average of 3.0° (range 0.7-5.0°), corresponding to 33% of the geometrically required adaptation.

Linear vestibuloocular reflex during motion along axes between nasooccipital and interaural

Linear vestibuloocular reflexes (LVORs) stabilize retinal images by producing eye movements to compensate for linear head motion. LVOR response characteristics depend upon gaze relative to the motion axis and binocular fixation distance. LVOR sensitivity during NO-axis motion increases as gaze eccentricity relative to the motion axis increases and as binocular fixation distance decreases. To fixate targets during forward head motion and rightward gaze, eyes must move to the right, but when looking left, the eyes must move to the left. In this study, LVORs were measured (binocular search coils) during 5.0 Hz horizontal motion along axes between and including NO and IA. This reorients head and otolith inputs relative to linear motion. We found that LVORs follow the same kinematics regardless of eye position in the head or head orientation relative to motion. Eye position information must be quickly and accurately integrated with otolith inputs to determine eye position (gaze) relative to linear head motion in space. The LVOR provides a behaviorally useful reflex for maintaining ocular fixation on visual targets during translation along any axis.

Spatial Orientation of VOR to Combined Vestibular Stimuli in Squirrel Monkeys

The interaction of angular and linear stimuli produces a complex alignment of spatial orientation and the VOR. This phenomenon was studied by measuring three dimensional eye movements in 6 squirrel monkeys during centrifugation in the dark. The axis of eye rotation was always aligned with gravity and with the spinal axis of the upright monkeys. The erect monkeys were oriented such that they were either facing toward the direction of motion or were facing away from the motion. Angular velocity trapezoids were utilized as the motion stimuli with a ramp acceleration of 10 degrees/s2 to a constant velocity of 200 degrees/s. This yields a final centripetal acceleration of 1 g. The orientation of centripetal acceleration dramatically altered the VOR by changing the axis of eye rotation, the peak value of slow phase eye velocity, and the time constant of per-rotary decay. The axis of eye rotation always tended to align with gravito-inertial force, the peak value of slow phase eye velocity was greater when the monkey faced the motion than when it faced away from the motion, and the time constant of decay was smaller when the monkey faced the motion than when it faced away from the motion. These findings were statistically significant (p less than 0.05) and were consistent across all monkeys. The data also indicate that the VOR may be separated into two reflexes, a linear reflex and a rotational reflex. The linear reflex decays as the axis of eye rotation aligns with gravito-inertial force (GIF). These results indicate that GIF is resolved into two components: one representing an internal estimate of linear acceleration and one representing an internal estimate of gravity.

Neural Geometry Revealed by Neurocomputer Analysis of Multi-Unit Recordings

A basic goal of neurocomputing is to identify, from biological neural nets, the mathematics intrinsic to neural net function. This is not only a basic research goal — but fundamental to electronic implementation to neurocomputers. To experimentally measure neural geometry, from up to ten cerebellar Purkinje cells of locomotory cats, multi-electrode recordings were obtained and the metric tensors of the functional manifold were calculated. The covariant metric tensor was calculated by cross-correlation analysis and the contravariant metric tensor was obtained by its Moore-Penrose inverse. For analysis of massively parallel data, a transputer-based neurocomputer was built. Results demonstrate the non-Euclidean geometry intrinsic to neural nets and its modifiability with learning.