Adaptation Of The Vestibulo-Ocular Reflex Research Paper

1. Vestibulo-Ocular Reflex

The reflex arc for the VOR is a three-neuron pathway connecting the primary vestibular afferents arising from the semicircular canals, to the VOR relay neurons in vestibular nuclei, and finally to motoneurons innervating the extraocular muscles. VOR relay neurons are either excitatory or inhibitory in their synaptic action, and induce either contraction or relaxation of the extraocular muscles via excitation or inhibition of motoneurons. There are a number of parallel pathways connecting the three semicircular canals (horizontal, anterior, and posterior) and the two otolith organs (saccule and utricle) in each labyrinth to the six extraocular muscles in each eye (medial and lateral rectus, superior and inferior rectus, superior and inferior oblique). These pathways operate in concert under conditions of free head movement, but in an experimental setup they are stimulated separately by giving yaw, pitch, roll or linear acceleration.

The horizontal VOR is tested by sinusoidal or velocity-step head rotation on the horizontal plane, and the VOR gain is measured as the ratio of the attained eye velocity to the applied head velocity. Sinusoidal rotation is convenient for measuring the gain and phase of the VOR separately, while velocity steps enable us to separate the VOR responses, which arise with different latencies, into their constituent components. Measurement of the VOR is performed in the dark or with the eyes closed while the whole body is rotating.

2. Vestibulo-Ocular Reflex Adaptation

Long-term visuovestibular mismatching (for days to months) can be imposed in a subject (either animal or human) with the aid of a pair of prism or lens goggles. Short-term mismatching (1–4 hours) is created by combined rotation of the turntable on which the subject is positioned and a screen representing the visual surroundings. Both wearing of Dove prism goggles, which reverses the right–left axis of the visual field, or miniaturizing lenses which reduce the visual field, and in-phase rotation of the turntable and screen in the same direction, cause adaptive reduction of the horizontal VOR gain. Likewise, both wearing of 2x magnifying lenses and out-of-phase rotation of the turntable and screen in opposite directions cause adaptive increase of the horizontal VOR gain. These adaptive changes in VOR gain act to reduce the retinal slips that occur during the VOR, which otherwise are augmented by the imposed visuovestibular mismatching.

A cross-modal adaptation occurs between the horizontal and vertical VORs; when yaw rotation on the horizontal turntable is combined with vertical optokinetic motion, the VOR measured in the dark acquires a vertical component even when the head is rotated only horizontally.

3. The Flocculus

In rabbits and cats, the flocculus has five to six major folia, while in rats it usually has only one folium. In classic anatomical studies of the monkey cerebellum, ten folia have been described in the flocculus, but recent studies on neuronal connectivity have revealed that the rostral five folia belong to the ventral paraflocculus and not to the flocculus.

A proportion of Purkinje cells in the flocculus of either side supply inhibitory synapses to both excitatory and inhibitory relay neurons in the VOR arc arising from the ipsilateral semicircular canals. Purkinje cells related to the horizontal VOR arc are localized to a narrow zone (H-zone, about 1 mm across) in the flocculus. The H-zone is flanked by other zones related to vertical (V-zone) or rotatory (R-zone) VORs, and a simple method to identify these zones in physiological experiments is to apply local electrical stimulation to these zones, which induces distinct horizontal, vertical, or rotatory eye movements.

These microzones in the flocculus are also characterized by their climbing fiber inputs relayed by the inferior olive. Purkinje cells in the H-zone receive climbing fiber signals encoding retinal errors, that is, slips of the retinal images on the horizontal plane. Purkinje cells reported to receive climbing fiber signals encoding retinal slips approximately in the vertical and rotatory directions, may be located in the Vand R-zones, respectively.

Branches of mossy fibers arising from the semicircular canals are distributed throughout the flocculus. Canal signals could also be relayed to the flocculus via 4–6 mm-long parallel fiber axons of granule cells. Yet, in the H-zone Purkinje cells, simple spike discharges are modulated specifically in relation to the horizontal head velocity (see below).

4. Flocculus Hypothesis For Vestibulo-Ocular Reflex Adaptation

Based upon the abovementioned relationships between the flocculus and the VOR, it is suggested that the H-zone originally receives mossy fiber information from all directions, and through learning by referring to climbing fiber signals, canal-plane-specific simple spike responses are established. It is further suggested that the established simple spike response patterns are updated daily by learning, whenever retinal errors are sensed through the climbing fibers.

Learning in the flocculus would occur at the synapses formed between the axons of granule cells relaying mossy fiber signals and Purkinje cells. This type of interaction was suggested in Marr’s and Albus’ theories of the cerebellar circuit proposed around 1970, later identified as long-term depression (LTD). LTD occurs in granule cell axon-to-Purkinje cell synapses, when these are conjointly activated with climbing fibers. These hypothetical views have been tested experimentally in recent decades, as reviewed below.

5. Deprivation Of Flocculus Function

That the flocculus plays a role in VOR adaptation has been confirmed by showing that impairment of flocculus function results in abolition of VOR adaptation. Previously, surgical ablation of and injection of toxic amino acids in the flocculus were used. More recently, based upon newly acquired knowledge about LTD, various pharmacological reagents and genetic means have been used to block LTD in the flocculus. Injection of hemoglobin into the subdural space over the flocculus in rabbits and monkeys abolished VOR adaptation, presumably due to the scavenging action of hemoglobin on nitric oxide (NO) that is required for the induction of LTD. Injection of an inhibitor of NO synthase into goldfish cerebellum inhibited adaptive increase of VOR gain. Mice transfected with the gene of a pseudosubstrate inhibitor of protein kinase C (PKC), did not exhibit VOR adaptation, as expected from the well-known requirement of PKC for LTD induction.

6. Behavior Of Flocculus Purkinje Cells

Whether the behavior of H-zone Purkinje cells is consistent with the flocculus hypothesis has also been examined by recording two types of spikes generated by the Purkinje cells: namely, simple spikes, reflecting their mossy fiber inputs, and complex spikes, evoked by climbing fiber inputs.

During sinusoidal head rotation, VOR relay neurons are driven by signals from the ipsilateral horizontal canal, which are modulated in phase with the head velocity (increased during ipsilateral rotation). In the H-zone, Purkinje cells exhibit modulation of simple spikes either in phase or out of phase (decreased during ipsilateral rotation). The in-phase-modulated inhibitory signals of Purkinje cells should depress VOR by canceling the effect of similarly modulated canal signals. On the other hand, the out-of-phase discharges of Purkinje cells should enhance VOR by adding inversely modulated inhibitory signals to the canal signals.

Under the condition that sustained visuovestibular mismatching enhances the VOR, the simple spike responses of H-zone Purkinje cells to head rotation became predominantly out of phase. Conversely, under the condition that sustained mismatching de- presses the VOR, the simple spike response became predominantly in phase. These changes in simple spike responses can be interpreted as being causally related to the VOR gain changes.

The two types of simple spike discharges in H-zone Purkinje cells may be induced by mossy fiber inputs from the horizontal canals of the two sides. Signals from the ipsilateral canal would drive Purkinje cells in phase, while signals from the contralateral canal would drive Purkinje cells out of phase. The shift of dominance of the type of simple spike modulation during VOR adaptation can be explained as being caused by LTD, depressing either the ipsilateral or contralateral canal inputs to the Purkinje cells. The climbing fiber discharge patterns observed in H-zone Purkinje cells are consistent with this notion.

7. Memory Site For Vestibulo-Ocular Reflex Adaptation

No doubt, the flocculus plays a crucial role in the induction of VOR adaptation. However, there is a dichotomy of opinions regarding its role in retention of the memory of VOR adaptation. To determine whether or not the memory of VOR adaptation is retained in the flocculus, studies have been carried out on the effects of blocking of flocculus functions— either surgically or functionally—after VOR adaptation develops. In goldfish, no less than 30 percent of the altered VOR gain was still maintained after ablation of the cerebellum, suggesting that the cerebellum is the major, but not the sole, site responsible for the maintenance of VOR adaptation. Microdialysis of lidocaine into goldfish cerebellum, however, abolished both the induction and short-term maintenance of VOR adaptation, suggesting that the memory for VOR adaptation is retained at least for a few hours in the cerebellum.

Another strategy to determine whether the memory underlying VOR adaptation is retained in the brainstem or in the cerebellum is to analyze the component eye movements elicited by velocity-step head stimuli at different latencies. The measurements in goldfish indicated that adaptive changes occurred in the earliest component of the eye movement elicited by a stimulus with a latency of 18 milliseconds, which could be mediated via the three-neuron VOR arc. However, the measurements in monkeys remain inconclusive.

8. Oculomotor Signals In The Flocculus

A complication in the flocculus hypothesis arises from the possible existence of mossy fibers that relay information regarding oculomotor activities to the flocculus. The existence of such mossy fibers has frequently been assumed in modeling the flocculus– VOR system (Lisberger and Sejnowski 1992).

However, it is difficult to accurately estimate such mossy fiber signals in isolation from the vestibular or retinal mossy fiber signals. In the previously used visual suppression method, the oculomotor signals were eliminated by training a monkey to fix its gaze on a spot moving with the head, and the resultant changes induced in the Purkinje cell behavior are assumed to represent the oculomotor signals. However, such gaze fixation is executed by central command signals, which would be a source of another mossy fiber input to the flocculus. Oculomotor signals in the flocculus have also been measured during nystagmic eye movements evoked by the stimulation of the optic tract. This measurement yielded a much smaller estimate of the oculomotor signals to the flocculus as compared with the visual fixation method.

It is notable that the simple spike responses of Hzone Purkinje cells to head rotation in the dark and those to screen rotation that evokes optokinetic eye movement responses varied independently of each other. Since oculomotor components should be shared between the two types of simple spike responses, this observation in rabbits favors the view that H-zone Purkinje cells receive few oculomotor signals.

The possibility so far overlooked is that learning forms simple spike responses reflecting oculomotor activities. Since retinal errors represent the deviations of actual eye movements from the ideal eye movements fully compensating for head movement, these should contain information regarding eye movements, which could be reflected in simple spike responses through the LTD mechanism.

Bibliography:

1. Ito M 1998 Cerebellar learning in the vestibulo-ocular reflex. Trends in Cognitive Science 2: 313–21
2. Ito M 2001 Cerebellar long-term depression: Characterisation, signal transduction and functional rates. Physiological Reviews 81: 1143–95
3. Lisberger S G, Sejnowski T J 1992 Motor learning in a recurrent network model based on the vestibulo-ocular reflex. Nature 360: 159–61