Cortical visual-vestibular interaction for spatial orientation and self-motion perception

Two major cortical functions of the vestibular system are the perception of verticality and the perception of self-motion. Whereas the perception of verticality relies mainly on otolith input, the perception of self-motion involves both otolith and semicircular canal input. These functions also rely on visual and somatosensory input. The three sensory systems provide us with redundant information about the position and motion of our body relative to external space. No primary vestibular cortex Animal studies have identified several distinct and separate areas of the parietal and temporal cortices that receive vestibular afferents. These include area 2v at the tip of the interparietal sulcus, area 3aV in the central sulcus, area 7 in the inferior parietal lobule, and the parieto-insular vestibular cortex [1,2] at the posterior end of the insular and the visual posterior sylvian area [3]. The multiplicity of representations of vestibular cortical areas identified in electrophysiological animal experiments raises doubts about the existence of a primary vestibular cortex similar to the visual or auditory cortex. Most vestibular functions are not exclusively vestibular. Microelectrode recordings from the so-called vestibular areas all demonstrate that the neurones are multisensory (they respond to vestibular, visual and somatosensory stimuli) and the cytoarchitectonic structure of these ‚vestibular cortical areas‚ is also more characteristic of a multisensory or sensorimotor rather than a primary (unimodal) sensory cortex. Functional imaging has enabled visualization of human homologues of cortical areas that have been associated with vestibular function in positron emission tomography (PET) studies during caloric vestibular stimulation [4] and in functional magnetic resonance imaging (MRI) studies during galvanic stimulation [5]. The parieto-insular vestibular cortex as defined by Grüsser et al. [1,2] in the monkey seems to represent the dominant cortical vestibular area, which in humans is activated by both semicircular canal (caloric) and otolith (galvanic) stimulation. Functional MRI has also revealed that the parieto-insular cortex is involved in mediating visual and ocular motor function during small-field optokinetic stimulation [6]. These findings support the view that there is multisensory interaction of the vestibular system, especially with the visual system. Room tilt illusion Room tilt illusions are transient upside-down vision or apparent 90º tilts of the visual scene. They can be induced in patients with central vestibular disorders by either an acute vestibulocerebellar lesion or cortical dysfunction. The two causative cortical regions are the parieto-occipital area and, rarely, the frontal lobe [7]. A lesion in these regions can also cause spatial hemineglect. Tiliket et al. [8] reported that sudden 90º room tilt illusions were elicited in patients with unilateral brainstem lesions after vestibular stimulation by off-vertical rotatory chair rotation. Furthermore, patients with peripheral bilateral vestibular failure may report transient room tilt illusions on awakening in the morning. The latter observation may be comparable to the reports of astronauts of occasional upside-down vision in microgravity [9]. The mechanism of room tilt illusions is largely unknown. The vestibular and the visual systems provide us with cues about vertical orientation in a three-dimensional coordinate system. The three major planes of action of the vestibular system are the frontal roll plane, the sagittal pitch plane, and the horizontal yaw plane. Because two different verticals - visual and vestibular - cannot be perceived at the same time, there must be a cortical mechanism that integrates visual-vestibular input (by aligning the pair of axes) to determine the current percept of a unique verticality. We believe that room tilt illusions are transient mismatches of the cortical visual and vestibular three-dimensional coordinate maps, which occur in 90º or 180º steps as the erroneous result of the attempted cortical match (Fig. 1) [10]. Perceived tilt in one plane causes the afflicted person to attribute upright to horizontal or even down. The visual scene, which itself contains empirical spatial visual cues for up, will then in turn ‚dominate and correct‚ spatial orientation. Vision will tell the vestibular system where upright is. Therefore, room tilt illusions are transient; you cannot perceive two verticals at once. The matching of two separate sensory three dimensional coordinate maps must be plastic in order to compensate for visual-vestibular tone imbalance or to adapt to unusual environments (microgravity). The plasticity of this visual-vestibular interaction has been best demonstrated when wearing reversing prisms [11].Figure 1: Room tilt illusionReciprocal inhibitory visual_vestibular interaction for self-motion perception The vestibular system - a sensor of head accelerations - cannot detect motion at constant velocity and thus requires supplementary visual information. Visual perception of self-motion induced by large-field optokinetic stimulation (circularvection, linearvection) is essential [12]. Vestibular stimuli invariably lead to the sensation of body motion. Stimuli of visual motion, however, can always have two perceptual interpretations: either self-motion or object-motion. The subject who observes moving stimuli may perceive either himself as being stationary in space (egocentric motion perception) or the actually moving surroundings as being stable while he is being moved (exocentric motion perception). Visual self-motion can be perceived while gazing at moving clouds or a train moving on the adjacent track in a train station. Vestibular information about motion is elicited only through acceleration or deceleration; it ceases when the cupulae within the semicircular canals or the otoliths have returned to their resting position during constant velocity. Our perception of self-motion during constant velocity car motion is completely dependent on optokinetically induced vection. In order to determine the unknown cortical visual- vestibular interaction during circularvection, we conducted a PET activation study on circularvection in human volunteers [13]. The PET images of activated cortical areas during optokinetic stimulation without circularvection were subtracted from those with circularvection. Circularvection was shown not only to activate bilaterally a medial parieto-occipital visual area separate from middle temporal/medial superior temporal motion-sensitive areas, but simultaneously to deactivate the parieto-insular vestibular cortex. This finding supports a new functional interpretation: reciprocal inhibitory visual-vestibular interaction as a basic sensorimotor mechanism for adequate self-motion perception [13]. Such a mechanism protects visual perception of self-motion from potential vestibular mismatches caused by involuntary head accelerations during locomotion. This mechanism allows the dominant sensorial weight during self-motion perception to shift from one sensory modality (visual or vestibular) to the other. Depending on the mode of stimulation, perception of self-motion is dominated by either vestibular input (head acceleration) or visual input (constant velocity circularvection), or both. Quantitative visual-vestibular interaction is not a simple but rather a complex process; it reflects not only the pattern of motion stimulation but also particular active postural and locomotor tasks. The mechanism of reciprocal inhibitory visual-vestibular interaction makes it possible to relate the dominant perception of self-motion to the actual input of one of the two sensory modalities so as to avoid perceptual ambiguity. As a functional consequence, the concurrent deactivation of the vestibular cortex during circularvection should decrease the vestibular system‚s sensitivity to head accelerations. This would make the perception of optokinetically induced circularvection more robust and largely insensitive to visual-vestibular mismatches occurring during involuntary head accelerations in spatial planes different from the main direction of locomotion or transportation. The actual horizontal direction and speed perceived during constant velocity of car motion are transduced only by the relative optic flow of the surroundings. Concurrent vertical vestibular stimulations caused by car motion and secondary involuntary head accelerations provide vestibular information that is inadequate or even misleading with respect to self-motion perception in the horizontal direction. It is desirable that they be suppressed by deactivation of the vestibular system [13]. This hypothesis is supported by earlier findings that thresholds for detecting vestibular body accelerations (vestibular system) are significantly increased during optokinetically induced circularvection (visual system) in a combined rotatory chair-drum system [14]. In light of these considerations, an earlier observation can be interpreted as the vestibulo-visual pendant for self-motion perception. In a PET study using caloric vestibular irrigation [15], activation of the vestibular cortex caused a significant, bilateral decrease in regional cerebral blood flow in the occipital visual cortex that covers Brodmann areas 17, 18 and 19. This finding suggested that deactivation of the visual cortex is beneficial to the organism during vestibular stimulation, because it suppresses visual motion input (e.g. distressing oscillopsia, owing to the retinal slip of the visual scene during vestibular nystagmus). In the same way that deactivation of the visual cortex largely protects the vestibular system from conflicting visual motion input, deactivation of the vestibular cortex prevents optokinetically induced circularvection from conflicting with vestibular input (Fig. 2).