Simulation of vestibular semicircular canal responses during righting movements of a freely falling cat

The righting maneuver of a freely falling cat was filmed at 1000 pictures per second, and the head position about the roll axis was digitized from each film frame using a graphics input tablet. The head angular velocity and acceleration were computed from the roll axis position trajectory. Head acceleration trajectories approximated two periods of a damped sinusoid at a frequency of 26 Hz. Head acceleration peak amplitudes exceeded 120,000 deg/s2. These trajectories were used as stimuli for the horizontal semicircular canals in a computer simulation of first-order afferent responses during the fall. Linear system afferent response dynamics, characterized in a previous study of the cat horizontal canal using pseudorandom rotations, provided the basis for linear predictions of falling cat afferent responses. Results showed predicted single afferent firing rates that exceeded physiological values; and variations in afferent sensitivities and phase were predicted among different neurons. Fast head movement information could be carried by ensemble populations of vestibular neurons, and a phase-locking encoding hypothesis is proposed which accomplishes this. Implications for central program versus peripheral vestibular feedback strategies for motor control during falling are presented and discussed.