What visuospatial perception has taught us about the pathophysiology of vestibular migraine

Abstract

Purpose of review:

A decade has passed since vestibular migraine (VM) was formally established as a clinical entity. During this time, VM has emerged amongst the most common cause of episodic vertigo. Like all forms of migraine, VM symptoms are most prominent during individual attacks, however many patients may also develop persistent symptoms that are less prominent and can still interfere with daily activities.

Recent findings:

Vestibular inputs are strongly multimodal, and because of extensive convergence with other sensory information, they do not result in a distinct conscious sensation. Here we review experimental evidence that supports VM symptoms are linked to multi-sensory mechanisms that control body motion and position in space.

Summary:

Multisensory integration is a key concept for understanding migraine. In this context, VM pathophysiology may involve multisensory processes critical for motion perception, spatial orientation, visuospatial attention, and, spatial awareness.

Introduction

The association between headache, vertigo, and visual disturbances has been documented for centuries, going back to the work of Greek physician Aretaeus [1]. A distinct clinical classification for these constellation of symptoms started to take shape from mid 1980’s using various terms such as “migraine-associated vertigo,” “migraine-associated dizziness,” “migraine-related vestibulopathy,” “migrainous vertigo” and “vestibular migraine” [2-6]. Among these terms, vestibular migraine (VM) has become the accepted terminology, with now a decade since it was formally established as a clinical entity by the consensus of the International Headache Society (IHS) and the International Classification of Vestibular Disorders (ICVD) of the Bárány Society [7**]. During this time, VM has emerged amongst the most common cause of episodic vertigo. This is not surprising considering that (i) migraine may affect up to one in three females and one in ten males at middle age [8], (ii) vestibular symptoms are likely under-reported among migraine sufferers [9], (iii) migraine features are common among patients affected by vestibular symptoms [10], and (iv) VM is a diagnosis primarily based on clinical history [11]. Clinically, VM patients frequently describe misperceptions of self or environmental motion, or abrupt feelings of disorientation, imbalance, or tilt. These complex symptoms, defined as dizziness or vertigo, typically arise without any peripheral vestibular dysfunction, and frequently manifest with heightened sensitivity to visual or vestibular stimulations, like how other sensory stimuli result in common migraine symptoms such as allodynia, photophobia or phonophobia [12]. According to the ICVD, vertigo is a sensation of self-motion when no self-motion is occurring, and dizziness is defined as a sensation of impaired spatial orientation without a distorted sense of motion [13]. Similar to all forms of migraine, VM symptoms are most prominent during individual attacks, however many patients may also develop persistent symptoms that are less prominent and can still interfere with daily activities [7,14,15*].

Migraine involves a mix of sensory symptoms and their intricate interactions through an important neuro-physiological process known as sensitization [12]. This refers to how potentiation of neural response to sensory stimuli and their deficient habituation may trigger symptoms in patients. Accordingly, exposure to one sensory stimulus can result in hypersensitivity that extends to other sensory stimuli, leading to a generalized sensory dysmodulation. Consistent with this mechanism, VM patients present with symptoms that strongly correlate with cutaneous, visual and vestibular hypersensitivities. Indeed, cutaneous allodynia has been specifically correlated with the presence of vertigo [16*]. Additionally, visual motion can trigger a strong sensation of self-motion in these patients. The extent of overt vigilance on bodily sensation such as increased visual dependence is also associated with heightened level of alertness and anxiety in these patients [17]. Such observations are notably in line with the notion that migraine pathology is closely tied to abnormalities in multisensory processing and integration [12,18].

The term multisensory integration refers to how the brain constructs and updates a coherent perception of the environment using available sensory information. In this process, vestibular inputs are strongly multimodal, and because of their extensive convergence with other sensory information, they do not result in a distinct conscious sensation [19,20]. Similarly, balance and its perception are dependent on visual, somatosensory, cerebellar, and vestibular system [21]. For this reason, it is more pertinent to approach VM symptoms by examining key functions such as motion perception or spatial orientation. In this context, a more pertinent approach to VM pathophysiology is to link vestibular symptoms with multisensory brain functions that are pivotal for the control of body motion in space such as motion perception, spatial orientation, visuospatial attention, or spatial awareness. Here we present an overview of these processes and highlight their potential implications for understanding the putative pathophysiological mechanisms that contribute to vestibular symptoms in VM patients.

Multisensory mechanisms in vestibular migraine

The brain constructs and continually updates an accurate representation of the world by integrating sensory cues such as vestibular, visual, or somatosensory inputs into high levels of perception and consciousness. Psychophysical studies suggest that in this process, the internal sensory estimates are weighted in proportion to the reliability of the incoming sensory signals [19,22,23]. This form of integration is generally considered beneficial across perceptual functions, as it can improve sensory precision and contribute to adaptive behavior. However, in situations when different sensory systems are in conflict, it may lead to misperception if the mismatch results in an overreliance on one or more sensory modalities. In keeping with this cross-sensory mechanism of interaction, patients with vestibular pathologies tend to become overly dependent on visual cues [24-27]. These patients frequently experience dizziness or vertigo in visually rich environments or with exposure to visual motion (e.g., supermarket aisles, busy train station, or fast-moving visual scenes). Such increased reliance on vision or ‘visual dependence’ can be related to a compensatory re-weighting of sensory modalities due to reduced reliability of vestibular inputs [28]. Despite normal peripheral vestibular function in VM patients, they also experience similar symptoms with signs of overreliance on vision (Figure 1). This clinical manifestation suggests that VM pathology may involve neural mechanisms of sensory integration for perception of motion and spatial orientation.

Figure 1:

Clinical presentation of vestibular migraine is not associated with peripheral vestibular dysfunction. The results of vestibulo-ocular evaluation are presented in three patients with and without migraine symptoms or vestibular deficits, The video head impulse testing (vHIT) is shown as a measure of semicircular canal function, and the video ocular counter-roll (vOCR) as a measure of otolith function [63]. The vHIT traces show superimposed velocities of the head (orange) and eye (green) during lateral head impulses, with normal gains considered as eye velocity/head velocity >0.8. The vOCR traces show ocular counter-roll (blue) during head tilt of 30° (orange) with normal values>4° (negative values indicate left and positive values indicate right). Bar graphs show the mean vOCR values for three tilts on each side. Patient 1 has severe VM (vestibular migraine) symptoms with intact vestibular function and normal vHIT and vOCR results. Patient 2 also has severe vestibular and migraine symptoms, but this patient has bilateral vestibular loss involving both canal and otolith functions. With vHIT, there is a severe canal dysfunction causing very low gains (~0.2 ) along with compensatory catch up saccades. The vOCR in this patient shows low responses (~2°) consistent with loss of otolith function. Patient 3 also has severe canal and otolith hypofunction (vHIT gains~0.02, vOCR~2°). This patient, however, reports only oscillopsia with activity and some postural imbalance without any migraine symptoms. These examples show the multimodal nature of vestibular symptoms that are influenced by multiple factors beyond just vestibular inputs.

It is widely accepted that migraine involves sensitization of the trigeminovascular pathways, however it is not clear whether such central sensitization also occurs at the same level in VM patients, or if it mainly involves vestibular pathways. Some reports of spontaneous nystagmus during VM attacks or prolonged duration of vestibulo-ocular responses in these patients suggest that vestibular hyperexcitability may play a significant role during VM attacks [29-31]. However, it is not known whether this sensitization originates from the brainstem, cerebral cortex, or if it is related to a top-down modulatory effect of the thalamocortical network on the brainstem. Regardless of the source of central sensitization, altered activity in the vestibular system can impact various perceptual functions because of the strong and extensive multimodal sensory convergence of vestibular inputs. The structural and functional changes within the temporo-parietal cortex in VM patients provide evidence towards the involvement of these higher order neural mechanisms– in particular those that contribute to sensory integration for visuospatial functions [32-34**]. VM patients were also found to have higher thalamic activity in response to caloric irrigation in comparison with migraine patients without aura or healthy controls [35]. The magnitude of thalamic activation was positively correlated with the frequency of migraine attacks in VM patients. In addition, bilateral thalami showed higher activity during vertigo attacks in VM patients [32]. These findings agree with the studies that show the role of the thalamus as a major relay center for processing of vestibular inputs and their integration with other sensory modalities [36-38**]. However, little is known about how these sensory processes and the higher-level neural networks could be affected in VM patients. Interestingly, many patients undergoing vestibular testing such as caloric stimulation typically complain of triggered migraine symptoms. This would be in line with the evidence suggesting that vestibular stimulation can act as a migraine trigger [39]. Similarly, vertigo is reported in nitroglycerin-triggered migraine attacks [40]. These findings support the notion of migraine being a disorder of sensory processing in which sensitization plays a crucial role [12].

Motion perception

A coherent perception of motion relies on integration of allocentric and egocentric sensory information. In this context, the extent of reliance on each sensory modality is crucial, particularly in situations where the visuospatial context is in conflict with the sensory information that encodes the position of the body in space [20]. When there is conflict between visual and vestibular cues, migraine patients often experience strong illusions of self motion due to enhanced motion sickness sensitivity [41,42]. In these situations, the interaction of visual and vestibular cues can be quantified as a sensation of vection (i.e., false perception of self-motion or vertigo). Previous studies have reported that unlike healthy controls, VM patients experienced vection more strongly as motion duration increased [43,44]. VM patients also had higher thresholds to detect motion or generate vestibular ocular response during angular rotation [45]. These contrasted with healthy individuals and patients who had migraine without vestibular symptoms, but not those who had peripheral vertigo due to benign paroxysmal positional vertigo (BPPV). Similar to VM, patients with BPPV had higher vestibular ocular and perceptual thresholds. Thus, the abnormal thresholds were not exclusive to vestibular migraine, and they could be also linked to dizziness from another dysfunction that increased ‘noise’ within the vestibular circuits. After exposure to visual motion, however, there was a marked dissociation between the vestibular ocular and perceptual thresholds only in VM patients- and not in those who had BPPV. This finding is consistent with the observations that suggest VM pathology involves altered interactions between visual and vestibular inputs.

Although VM patients had higher perceptual threshold with yaw rotation, other reports have found lower thresholds with rotation in the roll plane compared to controls [46-48]. This discrepancy can be related to the type of vestibular stimulus and how perceptual thresholds were measured across different studies. The lower motion threshold in VM patients was based on a detection task (i.e., am I moving?) [46-48], whereas the higher threshold was measured with a discrimination task (i.e. ‘in which direction am I moving?’) [45]. These tasks may interact differently with how motion was perceived by patients in each study (Figure 2). A false sensation of motion (e.g., due to vestibular noise) can lower the threshold to detect motion, but it can make it harder to correctly determine the direction of motion (i.e., a higher discrimination threshold). Consistent with this notion, patients with VM – and not peripheral vertigo (BPPV)–had significantly increased error rates when reporting motion direction. Such findings prompt more questions about the underlying cause of perceptual changes in vestibular migraine; are they mediated by processes like cortical hyperexcitability, causing lower perceptual thresholds, or do they result from central sensitization at lower levels, causing increased perceptual thresholds? These key questions require further investigation by future studies. Given that the central sensitizattion is linked to migraine chronification [49], if this is also a key element in vestibular symptoms, it could be expected that patients with chronic VM would show further alteration in their perceptual thresholds compared to those with recent onset symptoms.

Figure 2:

A theoretical drift diffusion model can account for the motion perception thresholds with different experimental protocols. The top panel shows the model for discrimination thresholds (i.e., which direction am I moving; to the right or left?). In VM patients, the dizziness (i.e., vestibular noise) may create uncertainty regarding the decision, culminating in a longer time to reach the decision threshold compared to controls (i.e., raised motion perception threshold). The bottom panel shows the model for detection thresholds (i.e., Am I moving or not?). Here, the higher noise level in VM patients compared to controls can cause a quicker decision time to determine whether there is motion or not (i.e., lower vestibular motion thresholds).

Visuospatial orientation

Spatial orientation is another example that requires multimodal processing of vestibular inputs. This is related to how the brain must account for changes in the body position to maintain a constant orientation of the surroundings – a key function that is referred to as orientation constancy [19]. Spatial orientation is a multisensory process and different from the perception of body position itself. Therefore, when it is measured with respect to a reference like the direction of gravity, it doesn’t follow the same pattern as the perception of the body position. This point illustrates that spatial perception is not simply determined by vestibular inputs, and additional sensory information regarding the head and body positions must also be accounted for by the brain. Spatial orientation can be evaluated using the subjective visual vertical (SVV) task, which requires alignment of a visual line to earth vertical in the absence of any visual reference cues [19].

SVV biases reflect internal neural estimates that are computed via multisensory integration of signals that encode the eye, head and body positions. In upright position, SVV estimates are usually accurate, as they match the perception of the head position. At small tilt angles, however, there is an overestimation with SVV errors in the opposite direction of the head tilt (less than 45°- known as the E effect), and at large tilts, there is an underestimation with SVV errors in the direction of the head tilt (more than 45°- known as the A-effect) (reviewed in [19]). VM patients were found to have high variability in SVV responses with the head in upright position [50-52]. They also had larger SVV biases in the opposite direction of the head tilt, indicating abnormal sensory integration associated with overestimation of the head tilt position [53]. The larger SVV biases were found in the same head tilt direction that vestibular symptoms were reported by these patients. These results align with lower motion detection threshold during roll motion in VM patients compared with those who had migraine without dizziness or healthy controls [46-48]. The effect of visual motion on perception of spatial orientation has also been examined in VM patients. These patients had larger SVV biases when exposed to visual motion compared to controls [25,54]. Taken together, these findings suggest that VM patients display enhanced perceptual sensitivity with roll motion or tilt due to altered weighting of sensory signals that encode the head and body positions.

Visuospatial attention

Migraine symptoms may affect a range of brain functions that involve the acquisition, organization, and use of sensory information from the surrounding environment. In this context, sensory hypersensitivity in migraine patients could be linked with attentional processes that differentiate salient stimuli from other sensory distractors. Theoretical frameworks and experimental data suggest these attentional processes are mainly composed of (i) alerting, which refers to the ability to remain vigilant of upcoming sensory signals, (ii) orientating, which refers to the ability to align attention with the salient source of sensory signals, and (iii) executive processing or resolving conflict, which refers to the ability to readily focus on a task and suppress peripheral distractors [55,56]. These attentional mechanisms serve to make the incoming sensory information more readily available for various sensorimotor functions. Both migraine and vestibular dysfunctions can distinctly modulate separate components of this attentional network [57**]. Migraine patients without vestibular symptoms displayed changes in alertness associated with increased vigilance. Patients with peripheral vertigo (BPPV) had longer reaction times when they had to attend orienting spatial cues. VM patients, however, could stay vigilant and orientate to salient spatial information, but they showed a deficit when they had to focus on the task-relevant information and disregard the irrelevant stimuli [57**]. These observations are in line with the clinical symptoms from VM patients that are frequently triggered by irrelevant visual stimuli in busy environments.

Visuospatial perception is a dynamic process during which attentional mechanisms serve to make incoming visual information more readily available for various sensorimotor functions. In this process, not all visual inputs may reach the conscious state to become perceptually available [58]. Such gating of visual inputs may alter awareness of a visual stimulus despite its unchanged physical characteristics. This can be seen during the so-called motion-induced blindness (MIB) task, when salient visual stimuli may spontaneously disappear from visual perception when presented against a moving visual background [59]. Such a mutistable perceptual phenomenon suggests that while attention mechanisms are important in detecting visual information, another key aspect is the conscious awareness of the visual stimuli, which reflects how salient visual information is processed by the brain. The magnitude of perceptual fluctuation was higher in both patients with VM and peripheral vertigo compared to healthy controls or migraine patients without vestibular symptoms [60**]. There was, however, a different relationship between these motion-induced perceptual fluctuations and daily visual symptoms in these patient groups. VM patients had a higher rate of perceptual fluctuation with more severe symptoms, whereas BPPV patients had a lower rate of perceptual fluctuation with more severe symptoms. These findings suggest that while altered visual conscious awareness can be part of the pathophysiology in VM patients, it may represent a compensatory mechanism related to reweighing of sensory inputs to compensate for vestibular symptoms in patients with BPPV.

Conclusions

Here we reviewed recent experimental data that supports the working hypothesis that migraine and VM patients have impaired multi-sensory brain mechanisms that are critical for perception of motion and visuospatial functions. VM patients experience disabling spatial misperceptions including unusual sensitivity to head motion or visual stimuli, or sudden feelings of imbalance or tilt. These symptoms often co-occur and the presence and severity of one symptom may impact others. Therefore, it would be inadequate to consider each symptom in isolation and the entire symptom complex must be considered together for better understanding of underlying mechanisms. In this process, because vestibular inputs are strongly multimodal, they do not result in a distinct conscious sensation. For this reason, it is more pertinent to approach VM symptoms by examining key functions such as motion perception, spatial orientation, visuospatial attention or spatial awareness, which are integral to the control of body motion and orientation which respect to the surroundings.

Recent findings have provided preliminary insights into multisensory processes involved in VM symptoms. These include reports of abnormal motion-detection thresholds during roll tilt or yaw rotation, or larger errors of spatial orientation during head tilt, which suggest the brain of migraineurs respond differently to sensory stimuli, even interictally [45-48,53]. VM patients displayed deficits in executive functions characterized by an inability to focus attentional resources and suppress peripheral distractors [57]. There was also altered visual awareness in VM patients with higher perceptual fluctuation during exposure to visual motion [60]. Taken together, these findings suggest that altered perception and awareness of body position and motion in VM may be mediated by a combination of changes in the attentional network and multisensory processing in these patients. In this context, vestibular migraine can be considered as a failure to gate and process sensory inputs, including not only somatosensory inputs but also visual and vestibular cues that encode the body position and motion with respect to the surroundings.

Building on the notion that migraine reflects a sensory processing deficit, a more global and anatomically grounded view is that vestibular migraine reflects an altered brain state as a consequence of dysfunctional changes within the brainstem and thalamocortical networks. Therefore, it is likely that a large complex network of interactions mediates the alteration of sensory processing in these patients. At the cortical level, the tempo-parietal junction (TPJ) is a critical hub for integration of multiple sensory modalities, and it has been implicated in various aspects of spatial orientation including visuo-spatial attention, heading perception, visual gravitational motion perception, sense of embodiment, self-localization, and egocentricity [19,36,61,62]. The structural and functional changes within the temporo-parietal cortex in VM patients provide evidence towards involvement of higher order neural mechanisms– and in particular those that contribute to sensory integration for visuospatial functions. Given the extensive connections between the vestibular system, thalamus, and temporoparietal cortex, it is imperative to probe the interaction between these higher- and lower-level regions to further our understanding of VM pathophysiology, and begin to understand why vestibular symptoms may become prominent in migraineurs.

Key points.

• Vestibular migraine can be considered as a failure to gate and process sensory inputs, including not only somatosensory inputs but also visual and vestibular cues that encode body position and motion with respect to the surrounding environment.

• Building on the notion that migraine reflects a sensory processing deficit, vestibular migraine may reflect an altered brain state as a consequence of dysfunctional changes within the brainstem and thalamocortical networks.

• Given the extensive connections between the vestibular system, thalamus, and the temporoparietal cortex, it is imperative to probe the interaction between these higher- and lower-level regions in order to further our understanding of why vestibular symptoms can be prominent or less so among migraineurs.