Two-dimensional spatiotemporal coding of linear acceleration in vestibular nuclei neurons

Response properties of vertical (VC) and horizontal (HC) canal/otolith-convergent vestibular nuclei neurons were studied in decerebrate rats during stimulation with sinusoidal linear accelerations (0.2-1.4 Hz) along different directions in the head horizontal plane. A novel characteristic of the majority of tested neurons was the nonzero response often elicited during stimulation along the "null" direction (i.e., the direction perpendicular to the maximum sensitivity vector, S_max). The tuning ratio (S_min gain/S_max gain), a measure of the two-dimensional spatial sensitivity, depended on stimulus frequency. For most vestibular nuclei neurons, the tuning ratio was small at the lowest stimulus frequencies and progressively increased with frequency. Specifically, HC neurons were characterized by a flat S_max gain and an approximately 10-fold increase of S_min gain per frequency decade. Thus, these neurons encode linear acceleration when stimulated along their maximum sensitivity direction, and the rate of change of linear acceleration (jerk) when stimulated along their minimum sensitivity direction. While the S_max vectors were distributed throughout the horizontal plane, the S_min vectors were concentrated mainly ipsilaterally with respect to head acceleration and clustered around the nasooccipital head axis. The properties of VC neurons were distinctly different from those of HC cells. The majority of VC cells showed decreasing S_max gains and small, relatively flat, S_min gains as a function of frequency. The S_max vectors were distributed ipsilaterally relative to the induced (apparent) head tilt. In type I anterior or posterior VC neurons, S_max vectors were clustered around the projection of the respective ipsilateral canal plane onto the horizontal head plane. These distinct spatial and temporal properties of HC and VC neurons during linear acceleration are compatible with the spatiotemporal organization of the horizontal and the vertical/torsional ocular responses, respectively, elicited in the rat during linear translation in the horizontal head plane. In addition, the data suggest a spatially and temporally specific and selective otolith/canal convergence. We propose that the central otolith system is organized in canal coordinates such that there is a close alignment between the plane of angular acceleration (canal) sensitivity and the plane of lin-ear acceleration (otolith) sensitivity in otolith/canal-convergent vestibular nuclei neurons.