Spatial Properties of Otolith Units Recorded in the Vestibular Nuclei

INTRODUCTION

Central otolith-recipient units, which are found largely in the lateral and descending vestibular nuclei (LVN, DVN), have a wide range of polarization vectors, amplitudes, and phase characteristics.^1–8 Some central otolith neurons with canal-convergent inputs have spatial characteristics that are invariant. Other central otolith units respond to stimulation in all planes during sinusoidal oscillation around a spatial horizontal axis.^1,2,6–8 The temporal phase of the response of these units varies as a function of head orientation. These units have been called spatiotemporal convergence (STC) cells.² It has been suggested that STC cells have convergent inputs from canal and otolith afferents with different spatial and temporal properties, but the specific canal input to these otolith units has not been demonstrated. Using a technique for identifying specific canal input,⁹ we analyzed the spatial characteristics central otolith neurons with and without STC characteristics in the alert monkey and identified the specific semicircular canals that provide input to these neurons.

METHODS

Conclusions about canal and otolith input were inferred from the following tests: animals were sinusoidally oscillated about a spatial vertical axis while statically pitched forward and backward to demonstrate the relation of unit activity to the lateral and/or vertical canals. (In this study, we considered that all convergent inputs are excitatory.) If a unit has afferent input only from the lateral semicircular canals, then it will be maximally modulated when an animal is pitched forward about 30°.^3,9 If the unit has afferent input only from the vertical canals, it will be maximally modulated when the animal is tilted about 60° backward. If the neuron has convergent input from horizontal and vertical canals, then maximal unit modulation will occur somewhere between these two planes, resulting in complex phase relations. Otolith input, related to dorsoventral linear acceleration, is reflected as a bias frequency, when the head is tilted forward and backward during sinusoidal oscillation about a spatial vertical axis.

To determine if a unit has vertical-canal input, the animal is sinusoidally oscillated about a spatial horizontal axis with the head in different orientations to the plane of oscillation. Modulation of unit activity is plotted as a function of head orientation and fit with a sine function to obtain gain and phase of the spatial response. Assuming that an increase of unit activity will occur only with ipsilateral canal activation, then excitatory input can be related to a particular vertical canal if the spatial phase of the response lies close (<22.5°) to the canal plane. During sinusoidal oscillation around a spatial horizontal axis, otolith input can mimic a vertical-canal response, but otolith sensitivity will be detected when the unit is tested with static head tilt at the same orientation.

RESULTS

Sixteen neurons, predominantly sensitive to head orientation relative to gravity, were recorded in the vicinity of the LVN and DVN in three cynomolgus monkeys (Macaca fascicularis). Neck proprioceptive input was tested by pressing on the neck muscles. Only units that did not have clear neck proprioceptive input and no relation to velocity storage were analyzed in this study. The resting discharge of these units varied from 11.8 to 60.6 imp/s (average 31.0 ± 15.8 imp*s⁻¹) (Table 1). The average coefficient of variation (SD/average frequency) ranged from 0.2 to 0.55 (average 0.36 ± 0.07). Therefore, the units were classified as irregular.

TABLE 1:

Summary Table of Central Otolith Neurons Characteristics

Convergent Inputs	Response-Vector Orientation			Resting Discharge	Sensitivity (imp*s−1g−1)		Monkey ID
	Static Stim.	Sine 0.25 Hz	Ovar	Ave+SD	Static Stim.	Sine 0.25 Hz	Unit Id
None	—	97°	≈64°	14.9 ± 1.8	—	47.3	M9357(07)
None	—	282°	296°	60.3 ± 9.1	—	33.3	M9304(68)
None	315°	321°	314°	32.1 ± 2.9	35.6	38.0	M9358(07)
c-PC	126°	127°	113°	80.5 ± 4.1	43.3	89.5	M9357(38)
c-PC	288°	261°	258°	53.8 ± 4.6	22.1	38.1	M9304(69)
i-AC	347°	358°	332°	36.3 ± 1.2	25.0	73.1	M9358(09)
i-PC	—	—	123°	34.6 ±8.8	—	—	M9304(22)
i-AC & I-PC	125°	89°	105°	17.7 ± 2.6	14.8	50.1	M9357(46)
c-AC & c-PC	—	283°	278°	45.3 ± 2.8	—	51.4	M9304(70)
Body muscles	—	—	51°	66.4 ± 11.4	—	—	M9304(33)
Spatiotemporal Convergence (STC) Cell
i-AC	67°	80°	50°	39.3 ± 3.4	18.4	31.3	M9357(51)
i-AC & i-PC	82°	89°	79°	40.4 ± 6.6	17.3	27.8	M9358(11)
i-AC & i-PC, c-LC	97°	112°	89°	24.5 ± 1.2	17.0	40.5	M9358(10)
bi-PC	—	181°	138°	77.2 ± 7.7	—	31.0	M9357(17)
bi-PC	—	186°	178°	16.2 ± 2.9	—	34.0	M9357(40)
bi-AC	346°	10°	348°	37.7 ± 1.5	17.8	37.4	M9357(19)

Units were tested with static tilts while the head was differently oriented relative to the spatial horizontal axis. To determine the orientation of the response vector, neuronal activity was plotted as a function of head orientation in tilt and fitted with a sinusoid (Table 1). During off-vertical-axis yaw rotation (OVAR), the unit response vectors led head orientation for both directions of rotation. The direction of the response vector was defined as the phase of the maximal response averaged for clockwise and counterclockwise rotation (Table 1). The orientation of the response vector based on static or sinusoidal stimulation around a spatial horizontal axis or based on OVAR, was about the same.

Ten otolith-related neurons were activated during sinusoidal rotation about a spatial vertical axis while the head was pitched forward or back from 0° to 90°. In nine units, it was due to convergent vertical canal, and in one to lateral-canal input.

When tested with sinusoidal rotation around a spatially fixed horizontal axis, the average unit sensitivity to pitching was 45.9 ± 17.3 imp*s⁻¹*g⁻¹. Ten units had stable temporal phase leads of about 30° (varying from 0° to 50°) during pitching in various head orientations (Fig. 1A). In contrast, temporal phases varied as a function of head orientation in six units (Fig. 1B). These units had STC characteristics. That is, their activity was in phase with head position in one plane and in phase with head velocity in the orthogonal plane. Four STC units had vertical-canal inputs. Activity of another two units did not modulate during sinusoidal oscillation around a spatial vertical axis, but modulated in phase with head velocity during sinusoidal pitch forward/backward. They probably had equal excitatory inputs from both posterior canals (Table 1).

FIGURE 1.

Otolith neurons modulated during sinusoidal oscillation around spatially fixed horizontal axis. Gain was defined as positive (A and B, top) if the phase of unit modulation was less then 90° re stimulus position. Head orientation in tilt was defined as: 0°—nose down; 90°—right ear down; 180°—nose up; 270—left ear down. (A) Unit with stable temporal phase (A, bottom). (B) Phase of modulation of STC unit varied as a function of head orientation (B, bottom). (C) Sinusoidal oscillation at some head orientations had clear modulation at double frequency, indicating saccular afferent convergent input to this neuron. Unit activation by forward–backward (D) and left–right (E) tilts. Increase in activity for each direction was linear. The slope of the regression lines represents sensitivity for that direction of tilt.

Four STC neurons that were tested during sinusoidal modulation about a spatial vertical axis had a bias change as a function of head tilt. In one, the bias activity was increased with tilt forward and decreased with tilt backward. In another three, biases were minimal with the animal upright, but increased when the animal was tilted forward and backward (Fig. 1D). These units are similar to otolith afferents described as +Z.^10,11 One of the units had an activity increase when it was tilted laterally to the left or right in the roll plane (Fig. 1E). This unit may have had convergent input from two +Z afferents, which lay in two orthogonal planes. Thus, activity of at least four of six STC units in our study had saccular-type activation.

During sinusoidal oscillation around a spatial horizontal axis, the head passed through the upright position twice during each cycle. Therefore, Z units should modulate at double the frequency of oscillation. This was observed only in a few cases (Fig. 1C). In Baker et al.² several STC units indeed had modulation at a double frequency. They noted that “a few cells responded to both direction of rotation in one or more planes. …” In the other neurons in our study, it is possible that another modality, like vertical-canal input had modified the unit response, so that double peak did not appear.

DISCUSSION

This study shows that central otolith neurons located in LVN and DVN receive a wide range of convergent inputs from the semicircular canals. Our sample of otolith-related neurons, in the alert monkey had similar distribution of response vectors as shown for decerebrate cats.^1–5 Using our identification technique, it was possible to identify the specific canal input to the spatially identified otolith neurons. There were units that received input from a single canal (5/16). In other units (7/16), the input came from two canals, located on one or both sides. Of the six STC cells, four that were adequately tested had saccular input. Thus, we have demonstrated that the STC property of some central otolith neurons could be due to convergent input from the saccular maculae and specific vertical canals. Although preliminary, this study shows that there are sufficient types of otolith/canal-convergent neurons to implement the behavioral responses required for orientation and compensation during movement in three dimensions.