Vertical nystagmus received limited attention in the early medical literature (3, 4) until 1921, when it was experimentally induced in animals by severing the anterior semicircular canal (SCC) and observed in clinical contexts such as profound bilateral visual loss, drug toxicity (e.g., barbiturates, quinine), multiple sclerosis, spasmus nutans, brainstem lesions (e.g., pontine and medullary tuberculomas and meningiomas), and encephalitis lethargica (5). While some authors emphasized the importance of distinguishing central from peripheral etiologies in cases of vertical or mixed nystagmus (6), Walsh (7) asserted that purely vertical nystagmus was a hallmark of central or cerebellar pathology. He described cases associated with Arnold-Chiari malformation, cerebellar ataxia, bilateral pyramidal tract signs, and hydrocephalus, reinforcing the central origin hypothesis.

Upbeat nystagmus (UBN) typically follows Alexander’s law, exhibiting maximal fast-phase amplitude in upward gaze. However, in some cases, it may paradoxically increase with downward gaze (12, 13) or even convert into downbeat nystagmus (DBN) during upward gaze (13). These atypical patterns have been associated with lesions of the interstitial nucleus of Cajal (INC), which makes unstable the vertical neural integrator responsible for gaze holding - an association supported by corresponding MRI findings (2).

Convergence has a variable effect on UBN, being able to increase or decrease its intensity up to its suppression or in some cases transforming it into a DBN (14–17). At same time, vestibular stimulation, such as head-shaking and vibration, may reverse nystagmus direction. The changes in nystagmus patterns due to convergence and vestibular stimulation are likely due to disruptions in circuits that process linear acceleration signals from the otoliths, which are essential for generating the correct eye movements during head translation (18). A fundamental role in this is played by the translational vestibulo-ocular reflex (t-VOR), which during movement adjusts compensatory eye movements based on orbital position and vergence angle depending on the target position. In this way, during translation forward or backward the t-VOR generates horizontal movements for targets to the sides, vertical movements for targets above or below, and convergence/divergence for targets straight ahead. Key brain areas involved in these computations include the medial and inferior vestibular nuclei in the spinal cord, which mediate the velocity storage mechanism (19, 20), and their projections to the cerebellar nodulus, which computes the translational components of head movements (21). Disruption of these pathways or velocity storage mechanisms may explain the change in direction of nystagmus in response to convergence and vestibular stimuli, respectively (22).

Changes in head position relative to space, and hence in macular input, can abolish (8, 23, 24) UBN, increase its slow-phase velocity (25) and change its direction, often transforming it into a DBN (26, 27). UBN often tends to disappear in the supine or prone position, where the effect of the gravitational vector on the maculae is different (8). In a patient with a focal hemorrhagic lesion of the left BC, anterior vermis and left anterior superior cerebellar hemisphere, UBN was also suppressed by a contralateral head tilt due to otolith-ocular reflex activation (23). In a patient affected with a lesion at the cervicomedullary junction due to multiple sclerosis, the spontaneous upward and counterclockwise nystagmus was suppressed by the prone position, tilting the head to either side while sitting, and turning the head to the right while in the supine position. Moreover, nystagmus direction was reversed by hanging the head straight and turning the head to the left while the patient was in the supine position (24). Change from UBN to DBN were observed in the prone position in a patient with a lesion of the BC and moving from upright to a head-hanging position a patient with cerebellar vermis atrophy (28).

Another very peculiar feature of UBN is its possible spontaneous transformation, in the course of the disease, into a hemi-seesaw (29, 30), horizontal, torsional and above all DBN (17, 22, 24, 28, 31, 32). A very interesting case of the above phenomenon is Wernicke’s encephalopathy. The initial thiamine deficiency damages the perihypoglossal nuclei, particularly the nucleus intercalatus of Staderini (SIN) and the nucleus of Roller (RN) more than the PMT nuclei, resulting in a downward slow-phase bias and UBN. Over time, recovery is more complete in the perihypoglossal than in the PMT nuclei, which fails to provide excitatory input to the cerebellar flocculus. The consequence is that the cerebellar flocculus no longer exerts inhibitory control over upward slow-phase pathways resulting in an upward slow-phase bias and hence in DBN (30). Obviously, also the contiguity of the areas responsible for the vestibular syndrome in the sagittal plane may also justify the change in nystagmus direction (33, 34). Overall, changes in nystagmus direction with convergence, head tilt, vestibular stimuli or simply over the course of the disease reflect the tightly interconnected network that maintains balance in the pitch plane.

The neuronal circuits responsible for upward and downward postural and ocular reflexes are distinct within the central nervous system. These specialized pathways underlie the generation of different types of vertical nystagmus. Specifically, damage to the circuit facilitating upward compensatory and downward anticompensatory eye movements can result in UBN, while lesions affecting the circuit that facilitates downward compensatory and upward anticompensatory responses can lead to downbeat nystagmus (DBN) (35).

The effectiveness of the circuits controlling upward and downward vestibulo-ocular reflexes is not equal. Actually, upward eye movement reflexes are stronger than downward eye movement. Support for this assertion is the observation that optokinetic nystagmus and optokinetic after nystagmus as well as nystagmus induced by vertical rotation (36, 37) predominate for upward direction. Further confirmation is provided by the fact that the time constant for DBN is about 15 s, which is like that of horizontal nystagmus, whereas for UBN nystagmus it is about 8 s.

One reason for this vertical reflex asymmetry may be due to the anatomy of the eye. In fact, at least in the cat, the center of mass of the eyeballs is located anterior to the center of rotation (38, 39) and therefore, with subject in an upright position and the head in line with the trunk, the force of gravity limits the vertical VOR upwards and favors the vertical VOR downwards, creating a clear and maximum imbalance between the two vertical eye movement systems (Figure 1). On the other hand, the load and viscosity of the eye alone may not necessitate such a pronounced upward reflex preponderance.

An alternative explanation considers that gaze stability is maintained not only by the vestibulo-ocular reflex (VOR) but also by the vestibulocollic and cervicocollic reflexes (40). In the case of head reflexes, the influence of gravity is significantly greater than that acting on the eyes, requiring a robust upward enhancement. Consequently, the observed upward bias in the ocular reflex may also serve to compensate for potential gain and phase deficits in the head reflex, thereby providing effective counteraction to gravitational forces.

In addition, the upward preponderance should be modulated depending on the position of the head re-gravity, so that an influence from the otolithic signals, sensing the orientation of the gravity vector, is required for inhibiting or enhancing the upward responses depending either on the direction of the gravity vector and on the load and viscosity in the motor system involved.

At the peripheral level, the upward preponderance in vertical eye movement reflexes is supported by the anatomical and physiological properties of the vertical semicircular canals (VSCCs) (41). VSCC activation occurs via an ampullofugal stimulus, with the receptors of the anterior SCCs being more effectively stimulated by downward rotations and the receptors of the posterior SCC (PSCCs) responding more strongly to upward rotations. Because the anterior SCCs (ASCCs) are more closely aligned with the sagittal plane than the posterior canals, the signals from vertical head rotations are inherently asymmetric, producing a stronger response during downward rotations.

The most important central pathway facilitating upward eye movements is the crossing ventral tegmental tract (CVTT). Additional areas potentially involved in such an asymmetry include a circuit consisting of Superior Vestibular Nucleus (SVN), the RN and the SIN, and a circuit in which the PMT are involved.

This complex crossing pathway, responsible for upward eye movement facilitation, originates in the upper pole of the SVN, courses predominantly in the pons, then in lower midbrain tegmentum and finally reaches the III (oculomotor) nucleus (42, 43).

Afferences to SVN come from the anterior semicircular canals, both directly both indirectly through the flocculus, via the interstitial nucleus of the vestibular nerve (2); the maculae, both directly and through the flocculus; the caudal medulla, in which RN and SIN ‘s receives a collateral pathway from the CVTT and projects to the flocculus, for plausible feedback control; the flocculus, which receives inputs from the caudal medulla, ASCCs, maculae and visual apparatus and projects to the SVN, ensuring control of the circuit (44).

The CVTT can be regarded as a true anti-gravitational pathway controlled by macular input (Figure 2) which constantly modulate the tonic activity of the oculomotor nucleus based on the orientation of the head and of the gravitational vector, counterbalancing gravitational forces in relation to the head and eye’s static position and correcting the eyeball’s vertical inertial asymmetry. When the head is in upright position, the gravitational vector facilitates the CVTT so that the upward eye responses are potentiated. In prone or supine positions, the macular input changes and diminishes its facilitatory effect in the CVTT, so that the eye positions balance is therefore guaranteed only by the activity of the pathways that run in the MLF (see below). Finally, in the vertical head-down position, the upward facilitatory effect of CVTT will be minimal or absent and the gravitational vector will tend to favor upward eye movement relative to the head. In absence of CVTT-like circuitry compensating for such an eye movement due to gravity, an upward slow phase and chin-beating nystagmus may occur (45–47). Moreover, by providing the eye elevator muscle motoneurons with precise tonic activity, the CVTT integrates the function of the excitatory MVN-MLF pathway which, in the absence of gravity, would theoretically be sufficient by itself to manage the vertical movements of the eye. To put it more clearly: the excitatory MVN-MLF pathway controls upward eye movements, calculating the eye velocity relative to the orbit in response to rotational and/or translational head movements. The CVTT adjusts these parameters based on the instantaneous gravitational vector, i.e., the spatial position of the head.

The existence of a neuronal circuit facilitating upwards eye movements has been confirmed by numerous experiments performed in hypogravity and hypergravity (48–50) and above all from the clinical findings.

The SVN is also controlled by a loop located within the dorsal caudal medulla, consisting mainly of the RN and the SIN, both belonging to the perihypoglossal nuclei (PHN), and the flocculus. For completeness, it is worth mentioning the nucleus prepositus hypoglossi (NPH), which is also part of the PHN but does not participate in this specific circuit.

PMT is a collective term used to refer to clusters of neurons scattered along the midline fibers tracts in the pons and medulla (66, 67) which receives inputs from all known premotor cell groups of the oculomotor system (52, 68–73) and project to the flocculus, paraflocculus, and vermis of the cerebellum (50, 66–68). The PMT contributes to neural integrator function by relaying eye movement signals to the vestibulocerebellum (67, 74). Particularly, the mid-medullary PMT-cell group in the monkey, the nucleus pararaphales (75, 76) receives vertical eye position signals from the interstitial nucleus of Cajal and therefore relays vertical eye position signals to the floccular complex. Thus, medullary lesions that affect the nucleus pararaphales may result in UBN (33, 77–81) and vertical gaze-evoked nystagmus (GEN). In addition, the PMT provide the cerebellum with a form of efference copy (or corollary discharge) of eye movement commands for the optimization of gaze control, including the fidelity of the neural integrator for eye movements or more long-term adaptive control of eye movements (74, 82).

For the sake of completeness, we also report the crossing path of the BC, which has long been considered responsible for UBN. This pathway arises from the SVN, runs rostrally in the caudal tegmentum, and decussate in the caudal midbrain before projecting projects to the III (oculomotor) nucleus, which excites the motoneurons of the superior rectus and inferior oblique muscles, bilaterally. For a long time, this pathway was thought to be involved in the excitatory control of upward eye movements in the rabbit (83), until it was shown that this pathway was most likely confused with the CVTT, since these tracts are very close to each other in the lower pontine tegmentum (38, 39), and this is valid also in humans (23, 84, 85). Clinical data would confirm that a lesion of the BC does not determine an up-beat nystagmus consequent to the imbalance due to the predominance of the tonic activity of the system controlling downward movements as one would expect (86) but rather a down beat GEN (87) or a positional downbeat nystagmus (88). Moreover, some cases of spontaneous UBN due to involvement of the region through which the BC passes are characterized by large median tumoral or hemorrhagic lesions, always associated with damage to the cerebellar vermis, which in itself may lead to UBN (2, 78).

To better understand the role of the pathway running through the medial longitudinal fasciculus (MLF) in the genesis of UBN, valuable insights can be gained from patients with lesions of this pathway, in whom high-acceleration head rotations reveal an asymmetrical vertical vestibulo-ocular reflex (VOR) gain deficit. Specifically, upward VOR gain is less severely affected than downward VOR gain (89, 90). This finding supports the notion that upward VOR signals are transmitted not only via the MLF but also through extra-MLF pathways (91).

Additional noteworthy observations in such patients include the absence of spontaneous vertical nystagmus and the presence of predominantly upbeat GEN (53, 89–91). The lack of spontaneous vertical nystagmus suggests a relatively symmetrical role of the MLF in vertical VOR control. Meanwhile, the presence of upbeat GEN confirms MLF damage, indicating that the lesion deprives the interstitial nucleus of Cajal—a key component of the vertical neural integrator—of vestibular input integrator (2).

Finally, the induction of UBN following an MLF lesion requires selective or predominant damage to the bilateral pathways responsible for the upward VOR, while sparing those involved in the downward VOR. This selective damage may explain why upbeat nystagmus is infrequently observed in cases of internuclear ophthalmoplegia (24).

Although UBN is typically associated with central vestibular or brainstem lesions, rare cases of peripheral origin have been documented. Ichimura and Itani (100) reported a case of persistent positional UBN in a patient with bilateral posterior canal benign paroxysmal positional vertigo (BPPV) due to canalolithiasis. The nystagmus was elicited during the transition from an upright seated to a straight supine position, characterized by a latency of approximately 2 s and a maximum duration of 110 s. The absence of neurological signs, normal brain imaging, and spontaneous resolution supported a peripheral etiology. Similar cases have been described in earlier literature (101, 102). More frequently, a mixed spontaneous nystagmus with an upbeat component is seen in peripheral vestibular dysfunction. For instance, Fetter and Dichgans documented such a case in superior vestibular neuritis (VN) (103–105). In a retrospective study by Ling et al. (106), 43 patients with UBN were reviewed, and peripheral vestibular disorders were identified in 14 (32.6%) of them. These included 10 cases of superior acute unilateral vestibulopathy (AUVP), 1 complete AUVP, 1 probable labyrinthine infarction, 1 isolated acute unilateral utricular vestibulopathy, and 1 probable Ménière’s disease. Therefore, while rare, peripheral causes of UBN must be considered. A comprehensive diagnostic approach - integrating clinical history, symptomatology, neurological and neuroimaging evaluation, and especially what we call “semeiological features with different diagnostic weight” (e.g., nystagmus direction and characteristics, smooth pursuit, saccadic eye movements, skew deviation) are essential for accurate localization and differentiation.

Table 2 reports key points for distinguish between central or peripheral origins of the UBN.