Multisensory control of gaze stabilization : from brainstem to bedside

Abstract

Without the continuous updates provided by the vestibular and visual systems, our world would appear blurry and unstable. In order to allow the retina to reliably record incoming light in a meaningful way, the vestibulo-ocular reflex (VOR) and optokinetic reflex (OKR) aim to stabilize our gaze on visual features. The VOR performs this function by causing the eyes to move in the opposite direction of the head, whereas the OKR allows the eyes to instinctively pursue a moving visual scene. Together, they make up the gaze-stabilizing eye movements available to us humans, and represent the origins of all vertebrate oculomotor control. The neurophysiological principles governing these responses appear to have been very well-conserved throughout evolution, although the basic mechanisms through which the VOR and OKR are integrated in the brainstem are still not fully mapped; in nominal conditions, gaze-stabilization will rely on seamless merging of visual and vestibular information at this level. While it is difficult to analyse the activity of such an integration in humans, tracking the VOR and OKR can offer valuable insights into how the brain responds to motion in different conditions. This thesis aims to further our understanding of gaze-stabilization by implementing a series of translational protocols which may be divided into three segments. 1) We will elucidate the basic neural principles and evolutionary origins of gaze-stabilization in the lamprey animal model, 2) investigate how eye movement responses reflects this visuovestibular integration in healthy subjects, and 3) test these findings in a clinical setting.

The basic science experiments featured a combination of behavioural and electrophysiological methodologies. In the first stage, video eye-tracking was used to monitor gaze-stabilizing eye movements in three dimensions. In this way we recorded both slow- and fast eye movements in all planes in what may be the phylogenetically oldest example of nystagmus. The next phase involved performing electrophysiological recordings to investigate the neural network that would allow for this primordial gaze-stabilization. As both visual and vestibular information is used to construct these eye movements, we constructed a rotating platform synchronized with two monitors for optokinetic stimuli. This platform was capable of holding an ex-vivo lamprey preparation while performing electrophysiological recordings of the extraocular muscles and vestibular nuclei. The next step was to establish whether the lamprey possess an OKR, which we were able to observe as eye muscle activity scaled with optokinetic velocities in the pitch and roll planes. Combining visual and vestibular stimulations in the tilting platform, we saw that the conjoined OKR-VOR exhibited clear additive properties, which favoured the vestibular modality as the velocity was increased. Lesioning experiments showed that pretectum is vital for relaying the optokinetic signal towards the OKR, whereas tectum downregulates its intensity. Curiously, there was no clear OKR in the yaw plane. We did however record clear compensatory eye movements during locomotion using a semi-intact preparation in the corresponding directions. As a way to map the trajectory of the signals underlying these responses we performed tracer injections of key structures in the brain. As a result, we were able to visualize the fundamental neural network of gaze-stabilization.

Implementing similar principles as those outlined above, we tested how healthy human gaze-stabilization behaves in the roll plane under various circumstances. Using a mechanized sled capable of full-body rotations coupled with visual stimulations on a projected screen, healthy subjects wore an eye-tracking device while exposed to different levels of visual clutter and movement accelerations. Roll plane VOR and OKR responses can be monitored in terms of the ocular torsion, which sees the eyes torque around their axes. Unlike translational gaze-stabilizers this eye movement lacks any meaningful somatic innervation, meaning that it offers a clearer reflection of the fundamental reflex. Throughout these studies we recorded increased gain in ocular torsion to intensified visual and vestibular variables. In addition, when comparing the torsional response for the VOR and OKR to the joint visuovestibular response it became clear that they reliably produced summative responses. This phenomenon was then used to quantify the relative influence of each sense during movements.

Vertical vergence is a known physiological response to a head tilt and associated with ocular torsion. When performing the previously outlined studies we found that such a divergence of the eyes may also be triggered by optokinetic rotations. The torsion-vergence ratio increased to visual clutter and decreased with an upregulated acceleration of the stimulus motion. In order to ensure that the vergence was not secondary to torsion, we tested how it responded to visual information by having subjects view a rotating scene binocularly as well as monocularly. Results showed the vergence amplitude was increased during monocular viewing, suggesting that the response is suppressed during nominal binocular viewing. Furthermore, as postural control also involves similar multisensory modalities as gaze-stabilization, the relationship between postural sway and the OKR was tested. This involved a procedure where subjects watched a rotational optokinetic scene while standing on a balance board. Torsional gain was positively correlated with postural sway, as well as with the sympathetic response indicated by an increased pupillary dilation. Having established that the OKR-VOR interaction could be used to infer their relative importance during a head roll, we implemented our visuovestibular protocol in a pharmaceutical study testing the effects of Meclizine. Antihistamines like these are widely used as anti-emetic drugs, and are primarily attributed with reaching their desired effect by inhibiting the vestibular system. Contrary to this hypothesis, the torsional response showed increased vestibular influences during high velocities, although a converse inhibiting effect under nominal visuovestibular conditions at low accelerations. These findings may reflect the anti-emetic’s inhibiting effect on the central integration of the sensory information.

Through its translational approach to gaze-stabilization, the studies in this thesis offer new insights into the neural activity that govern eye movement control. The results presented above can be further contextualized in relation to the overarching aims of this work. 1) As the lamprey exhibiting OKR, VOR and locomotion-supported eye movements, it is clear that gaze-stabilization was well-developed already at the dawn of vertebrate evolution. The presence of nystagmus indicates that the neural mechanism for goal-oriented saccades was present at this time, and that it likely operates through tectal downregulation of the OKR. Pretectum has been shown to act as the first integrator of visual motion in several vertebrate species, which holds true also for the lamprey. In addition, this thesis shows that a VOR-OKR interaction can be maintained without cerebellar or cortical influences. The fundamental network instead relies completely on a few subcortical structures. 2) Our subsequent studies in human subjects revealed even more robust visuovestibular integration, showing that torsional gain may be used to test the influence of the sensory modalities in any given scenario. Our studies on the rotational OKR reveal that vergence is an intrinsic part of the response, and as it suppressed by binocular vision it is tempting to suggest it may reflect a visual activation of the vestibular nuclei. This visuo-vestibular pathway could also explain the correlation between visual clutter, postural sway and increased sympathetic stress, as it would allow reflexive postural strategies without conscious interference. 3) We implemented the aforementioned visuovestibular protocols to evaluate the effects of Meclizine, which suggested that it may be through central inhibitory effects that its anti-emetic effects are achieved. This finding is in line with well-established notion that 1st generation antihistamines are more efficient due to their properties on the central nervous system, being associated with stronger side-effects in the form of general sedation.

In conclusion, this thesis approaches gaze-stabilization from a broad perspective, shining new light on the neural network as well as identifying new features of the human eye movement response. Having established the fundamental principles that govern the integration of these reflexes, our findings open up for investigating the next order of integration, as several neural structures influence the gaze-stabilizing responses. The somatosensory influences towards this process remain poorly known, and future studies may benefit from mapping how the signal for locomotion-supported eye-movements is incorporated into the system. While we have indicated how torsion and vergence may hold utility, additional studies are needed to formalize the measuring of such processes for clinical use. Altogether, these findings could offer a valuable approach to evaluating sensory deficits, and aid in developing protocols for personalized medicine and rehabilitation.