Trends in

Neurosciences
Review

Eye movements in Parkinson’s disease: from

neurophysiological mechanisms to diagnostic
tools
Chrystalina A. Antoniades1,3,* and Miriam Spering2,3,*

Movement disorders such as Parkinson’s disease (PD) impact oculomotor Highlights

function – the ability to move the eyes accurately and purposefully to serve a mul- Humans use different types of eye
titude of sensory, cognitive, and secondary motor tasks. Decades of neurophysio- movements to look at points of interest
in the visual scene, to optimize visual
logical research in monkeys and behavioral studies in humans have characterized
processing, and to guide hand and
the neural basis of healthy oculomotor control. This review links eye movement ab- body movements. Eye movements
normalities in persons living with PD to the underlying neurophysiological mecha- have the potential to reveal sensorimotor
nisms and pathways. Building on this foundation, we highlight recent progress in and cognitive processes in healthy and
pathological brains.
using eye movements to gauge symptom severity, assess treatment effects, and
serve as potential precision biomarkers. We conclude that whereas eye movements Eye movement abnormalities are
provide insights into PD mechanisms, based on current evidence they appear to common in Parkinson’s disease (PD)
lack sufficient sensitivity and specificity to serve as a standalone diagnostic tool. and can provide clues as to the identity
of neuronal populations, mechanisms,
Their full potential may be realized when combined with other disease indicators and pathways that contribute to disease
in big datasets. processes.

Eye movements may aid biomarker-

based diagnosis and quantification of
PD when combined with other biomark-
Eye movements as indicators of the brain pathophysiology in PD ers, although they currently appear to
Global increases in life expectancy and industrialization are associated with higher prevalence of lack sensitivity and specificity to serve
neurodegenerative conditions such as PD and related disorders [1]. PD is characterized by the as a standalone diagnostic tool.
motor symptoms rest tremor, rigidity (see Glossary), and bradykinesia, as well as gait and
Therapeutic approaches such as deep
posture problems, typically manifesting on one side of the body at the initial stages of the disease brain stimulation, in addition to their clini-
[2,3]. Nonmotor clinical features such as sleep disorders, gastrointestinal dysfunction, autonomic cal utility, offer a unique opportunity to
dysfunction, olfactory deficits, and psychiatric disorders can also occur and significantly deterio- systematically probe the brain networks
rate quality of life [4,5]. that underlie human oculomotor control.

Despite diagnostic criteria and guidelines [5], establishing an accurate and early PD diagnosis can
sometimes prove difficult. Adding to the multifactorial nature of the disease is the complication 1
Nuffield Department of Clinical
that hallmark motor symptoms develop at different time points and are often preceded by less Neurosciences, Medical Sciences
Division, University of Oxford, Oxford,
specific non-motor symptoms during an early, prodromal stage [4,5]. Moreover, several subtypes
UK
of the disease exist, emphasizing the importance of individualized diagnosis and treatment [6]. 2
Department of Ophthalmology & Visual
Over the past few years, there has been an increasing effort to characterize atypical Parkinso- Sciences and Djavad Mowafaghian
Center for Brain Health, University of
nian disorders. Even though new criteria have been proposed to improve diagnostic accuracy
British Columbia, Vancouver, Canada
[7], this group of syndromes can often be mistaken for PD, especially in their early stages. 3
These authors contributed equally to
this work.
Characteristic eye movement abnormalities [8–10] are commonly observed in PD and Parkinsonian
disorders (Box 1). Across neurological disorders, eye movements can help identify candidate
populations of neurons and brain pathophysiological processes involved in the neurological symp- *Correspondence:
[email protected]
toms [11,12], even in prodromal stages [13]. In PD, eye movements can provide basic research (C.A. Antoniades) and
insights into how specific cognitive and executive processes are impacted, and in clinical research [email protected] (M. Spering).

Trends in Neurosciences, January 2024, Vol. 47, No. 1 https://doi.org/10.1016/j.tins.2023.11.001 71

© 2023 Elsevier Ltd. All rights reserved.
 Trends in Neurosciences

Box 1. Hallmark oculomotor deficits in PD

Glossary
Humans use a combination of different eye movements, such as saccades and smooth pursuit, to shift or stabilize gaze
Antisaccades: a saccade away from
and to enhance focus on objects of interest in the visual environment. Saccades are typically examined for their latency,
(often in the opposite direction of) a target.
velocity, and amplitude, and whether they are made unwantedly and intrude fixation. They can be triggered externally by a
Antisaccade task: a specialized
sensory stimulus (e.g., visually guided saccades), or internally by instruction alone (e.g., self-paced saccades). Various ex-
experimental laboratory protocol in
perimental protocols have assessed the cognitive ability to predict, remember (e.g., memory-guided saccades), or
which observers are asked to look in the
suppress a target (e.g., antisaccades). In PD patients, saccades tend to be inaccurate and hypometric (fall short of their
direction opposite to a visual target.
target). With disease progression, saccadic latency may increase and velocity decrease [101]. Deficits are generally worse
Basal ganglia: group of subcortical
for larger, and for self-paced and memory-guided saccades. Patients make frequent errors in the form of early, inadvertent
nuclei responsible for motor control and
saccades toward a memorized target or the visual stimulus in the antisaccade task [102].
associated functions; disruption of the
basal ganglia network is associated with
Microsaccades are small, rapid eye movements that occur when a person is attempting to fixate their gaze on a stationary
symptoms of Parkinsonian disorders.
object. They are typically less than one degree of visual angle in amplitude and can occur at a rate of several per second.
Bradykinesia: slowness of movement
Microsaccades in PD can be altered, with changes in both the frequency and amplitude of these movements [41,42].
with a decrease in amplitude or speed as
Square-wave jerks, also present in healthy older people, may interrupt fixation as well [103]. Reports of ocular tremor
the movement is continued.
are controversial, and observed abnormalities might be compensatory eye oscillations resulting from head tremor rather
Deep brain stimulation (DBS):
than a fundamental feature of PD [104].
surgical implantation of electrodes to
stimulate target areas in the brain. In PD,
Smooth pursuit eye movements are examined for their latency and velocity relative to the target (velocity gain) and
DBS of the subthalamic nucleus and
catch-up saccades made to reduce the eyes’ position and velocity error relative to the target. PD patients can show
other brain areas may be conducted to
preserved pursuit, although pursuit initiation may be delayed and velocity gain reduced, owing to hypometria in catch-
relieve symptoms and enable a
up saccades. Pursuit deficits are more apparent in higher-level tasks involving tracking a remembered target or using visual
decrease in dosage of antiparkinsonian
cues for target selection [52]. Saccade, fixation, and pursuit deficits in PD patients reflect dysfunction in underlying neural
medication.
networks required to generate appropriate eye movement responses, including several interconnecting pathways that
Fixation: periods during which the eyes
involve the basal ganglia.
only move within a small spatial range
(microsaccades and drift).
Vergence abnormalities are common in PD [66–69] and might be linked to vision problems in PD, such as blur, double
Frontal eye field (FEF): a region in the
vision [67], and convergence insufficiency [65]. In general, eye movements are tightly linked to visual perception
frontal cortex involved in the control of
[105,106], and hence, oculomotor abnormalities might influence visual deficits reciprocally. Yet only very few studies
saccades and smooth pursuit eye
have simultaneously assessed oculomotor and visual deficits in PD in the same task (see Outstanding questions).
movements, including eye movement
inhibition and monitoring.
Gaze-evoked nystagmus: repetitive
studies, eye movements have emerged as a promising objective and non-invasive tool to charac-
and involuntary oscillations or to-and-fro
terize motor and cognitive dysfunction [14–16]. However, whether there is sufficient evidence movements of the eyes that are triggered
to date to reliably diagnose and differentiate between PD syndromes based on eye movement when the eyes are held in an eccentric
measurements is unclear. position (not straight ahead).
Ocular tremor: constant, involuntary
jitter of the eyes at a high frequency and
In this review article, we present recent advances in using eye movements to quantify PD. We first low amplitude.
link oculomotor deficits observed in PD to underlying neurophysiological mechanisms and path- Parkinsonian disorders: a group of
ways that had been identified in foundational studies of eye movement control in human and non- akinetic-rigid movement disorders that
includes PD, corticobasal syndrome,
human primates. We then critically evaluate the evidence for using eye movements as criteria for
multiple system atrophy, vascular
diagnosis and as tools to assess symptom progression and treatment outcomes in PD. parkinsonism, dementia with Lewy
bodies, and progressive supranuclear
Neural basis of healthy and abnormal eye movements palsy.
Pupil light reflex: in response to light,
Utilizing eye movements as an objective approach to quantifying disease capitalizes on several the pupil constricts, and in darkness, it
unique features of eye movements. People have a natural and spontaneous ability to generate dilates.
eye movements and can participate in eye movement assessment tasks with no or only minimal Rigidity: increased muscle tone.
instruction [17]. Eye movements are continuous and provide fine-grained spatial and temporal Saccades: rapid, ballistic movements
made by both eyes from one point in
information about sensory and cognitive processes, including information participants are not space to another, interspersed by
consciously aware of [18]. Eye tracking technology is now highly advanced and cost-effective; periods of fixation.
it can be portable and be used at the bedside. Different types of eye movements (Box 1) are Smooth pursuit eye movements:
slow, continuous responses of the eyes
flexibly deployed in various sensory stimulus- and task-settings, and the ability to execute and
to moving objects.
control these movements can be tested with relatively simple and noninvasive methods. Square-wave jerks: involuntary sac-
cades from and back to a fixation point.
As outlined in Box 1, eye movement abnormalities in PD occur across the eye movement spectrum. Substantia nigra pars reticulata
(SNr): a midbrain area implicated in PD
Yet, most studies have focused on saccades, presumably because they are the easiest to measure
and involved in the initiation of voluntary
at the bedside. Figure 1 shows a simplified and schematic overview of selected cortical brain areas

72 Trends in Neurosciences, January 2024, Vol. 47, No. 1

Trends in Neurosciences

movements and the suppression of

involuntary ones.
Subthalamic nucleus (STN): part of
the basal ganglia system. The STN is a
primary target site for deep brain
stimulation in PD.
SEF Superior colliculus (SC): a layered
nucleus in the midbrain involved in target
FEF selection and the initiation of eye and
PEF other movements, as well as the
integration of inputs from different
DLPFC sensory modalities.
Tauopathies: a group of
CN neurodegenerative disorders
SNr characterized by abnormal aggregation
of tau protein in the brain.
STN SC Vergence: disjunctive movements of
the eyes made to reduce visual blur and
achieve fusion of an object’s image for
binocular vision. Vergence eye
movements facilitate depth perception.
Frontal eye fields Parietal eye fields
FEF/SEF/DLPFC LIP/PEF

Subthalamic
nucleus

Caudate Hyper-
D1 D2 GPe direct
Indirect
Direct Substantia nigra
pars reticulata

Superior
colliculus
Excitation
To brainstem
Inhibition saccade generator

Trends in Neurosciences

Figure 1. Brain areas involved in the generation of saccades. Brain regions are illustrated on a schematic of the human
brain. The location of brain areas is approximate, and only selected areas and connections are shown. In the network
schematic, green arrows indicate excitatory connections (glutamate-mediated), and blue connections indicate inhibitory
connections (GABA-mediated). The SNr is the inhibitory gateway to the SC, and it can both promote and inhibit saccades.
Abbreviations: CN, caudate nucleus; DLPFC, dorsolateral prefrontal cortex; FEF, frontal eye field; GPe, external segment of
the globus pallidus; LIP, lateral intraparietal cortex; PEF, parietal eye field; SC, superior colliculus; SEF, supplementary eye
field; SNr, substantia nigra pars reticulata; STN, subthalamic nucleus. Figure inspired by Figure 7–13 in [19].

for saccade control. Many of the same areas are involved in the control of pursuit and small-scale
saccades (microsaccades) as well. Saccades are promoted and inhibited via projections to the su-
perior colliculus (SC) [19]. Whereas frontal and parietal eye fields have excitatory (glutamatergic)
connections to SC (Figure 1, green arrows), a second pathway (Figure 1, blue lines) links frontal
and parietal areas to the SC through the basal ganglia and particularly, the substantia nigra
pars reticulata (SNr). The SNr tonically inhibits the SC via GABAergic projections, thus inhibiting
saccades. At the same time, inhibition of the SNr itself can promote saccades. The SNr receives
inputs from at least three different pathways, which have all been implicated in PD: a direct pathway
with GABAergic projections through the caudate nucleus that promotes saccades, an indirect

Trends in Neurosciences, January 2024, Vol. 47, No. 1 73

 Trends in Neurosciences

pathway with GABAergic projections through the external segment of the globus pallidus (GPe) that
inhibits saccades, and a hyperdirect pathway with excitatory input directly from the subthalamic
nucleus (STN), cancelling saccades. Together, this system forms an inhibitory gateway to the
SC, potentially involved in reward-modulated and motivated behavior [20,21], and in deciding
whether to initiate a saccade [22,23]. We will return to elements of this pathway in our discussion
of antisaccade behavior, as well as the effects of deep brain stimulation (DBS) in PD.

In the human and non-human primate literature, the inhibitory control of saccades has been
probed with countermanding (stop signal), go/no-go and antisaccade tasks. In countermanding
or stop signal tasks, observers execute a speeded response (e.g., button press or saccade) to a go
signal, such as the disappearance of a fixation spot. In a subset of trials, participants must abort
that response when a stop signal (e.g., reappearance of fixation spot) is shown. This task probes
the initiation and cancellation of planned movements, and the monitoring of movement conse-
quences, subserved by the distinct networks in the frontal cortex [24,25] shown in Figure 1. Be-
cause of the tasks’ sensitivity to measuring inhibitory and impulse control – executive functions
that can be impacted in diseases affecting the corticobasal ganglia pathways [26] – antisaccades
have gained wide recognition for use in psychiatric and neurological populations [27]. Relatedly, the
antisaccade task (Box 2) has emerged as a precise means of quantifying neurological function
[28,29], and an internationally standardized protocol has allowed direct comparisons of results
across experimental settings and laboratories [30].

A hallmark of impaired inhibitory control in the antisaccade task is a significant increase in directional
errors. In addition, two key findings, related to inhibitory dysfunction, are commonly observed in PD:
the first, prolonged saccadic latency or reaction time, results from increased inhibition of the SC by
the SNr. The second finding is a tendency to make early, premature saccades in stop-signal tasks,
delayed reaction time tasks, or antisaccade tasks. This effect could relate to an independent process
that bypasses the SC-basal ganglia route [31]. It is important to note that even though PD patients,
as compared to healthy controls, show increased error rate, longer latency, and more premature
saccades in the antisaccade task, this behavior is as variable as the disease itself, as illustrated in
Figure 2, and is not specific to PD but also occurs in other neurodegenerative diseases [32].

Overall, the antisaccade task and similar protocols are useful because they not only mark PD
symptoms but can also shed light on the underlying brain pathophysiology. In the context of

Box 2. The antisaccade task

Saccades can either be elicited reflexively, by the sudden onset of a visual event or a loud sound, or voluntarily, by the desire to
look at a particular object or location. Similarly, individuals can voluntarily suppress the reflexive urge to make a saccade. In the
antisaccade task, the observer must suppress an automatic saccade to a suddenly appearing peripheral target (sometimes
termed the prosaccade) and instead look in the opposite direction (the antisaccade). This task thus requires (i) the initiation
of a reflexive saccade to the peripheral target; (ii) the subsequent inhibition of this saccade; and (iii) the initiation of a voluntary
antisaccade in the opposite direction. Such flexible control of action – orienting and suppressing a movement depending on
task and situational demands – characterizes executive control.

Neurophysiological studies in macaque monkeys have revealed the circuitry underlying successful saccade inhibition and
reorienting in the antisaccade task [107], making antisaccades highly serviceable for the investigation of diseases affecting
volitional, executive control. The examination of directional errors – commonly a short-latency prosaccade followed by a
corrective antisaccade – in conjunction with single-neuron recordings in saccade-control brain areas revealed a key role of
the SC and FEF in this task. The firing rate of SC [108] and FEF [109] saccade neurons is correlated with direction errors
as well as prolonged latency, indicating that lack of inhibitory input to these areas might underlie poor performance in this
task. A likely main source of such an inhibitory signal to the SC is the SNr (see Figure 1 in main text) in the basal ganglia
[110,111]. Congruent with findings of reduced dopamine levels in the basal ganglia and associated depletion of dopamine levels
in downstream areas such as the SC and brainstem [112], antisaccade performance is significantly impaired in PD patients.

74 Trends in Neurosciences, January 2024, Vol. 47, No. 1

Trends in Neurosciences

(A) (B) Figure 2. The antisaccade task,

and interindividual variability in task

Vertical position
Fixation
performance in Parkinson’s disease
Target cue (PD). (A) Typical trial timeline in a
standardized antisaccade task protocol
Saccade
Incorrect Correct [30]. Following a fixation period of
prosaccade antisaccade randomized duration, one target location
is cued. Cue onset indicates that a
(C) (D) saccade can be initiated (here shown
without a gap between fixation offset and
Vertical position

Vertical position

cue onset). (B–D) Individual 2D eye

position traces for three PD patients (data
from [50]), revealing large heterogeneity in
antisaccade deficits. Whereas the patient
shown in panel B performed the task
comparably to healthy individuals and only
Horizontal position Horizontal position made two incorrect prosaccades,
Trends in Neurosciences directional errors were more frequent in
the other two patients, who also showed
large vertical deviations.

assessing impulsivity, the antisaccade task has gained particular importance in examining effects
of DBS as a treatment for PD, as we discuss in the following sections.

Another saccade protocol that consistently reveals deficits in PD patients is the memory-guided
saccade task. In this task, a saccade is expected to be performed toward the location of a
remembered peripheral-onset stimulus after a brief delay. PD patients typically display accuracy
abnormalities (hypometria) [33,34]. Notably, these deficits are not unique to the oculomotor
system but occur in memory-guided pointing movements as well [35]. Although working memory
performance may be impaired in PD, hypometric memory-guided movements are believed to
result from a deficient memory-to-motor signal transformation process, rather than from cognitive
deficits, because manipulations of memory load do not further impair saccades [34,35]. Memory-
guided saccade abnormalities do not seem to respond to levodopa treatment [36] but may
improve with DBS (see below). These higher-order saccade deficits point at the involvement of
brain networks that include the basal ganglia and regions in the prefrontal cortex that support
working memory functions [37].

Saccades fall along an amplitude continuum from large to small [38,39], and microsaccades
are controlled by some of the same brain areas and mechanisms that also underlie saccades [40]
(Box 1). In PD, fixational eye movements such as microsaccades are often abnormal [41,42], although
disturbances are nuanced and patient-specific [43] and sometimes reported in healthy age-matched
controls as well. In general, PD patients and those with progressive supranuclear palsy (PSP) show an
increase in involuntary saccades during fixation (also termed saccadic intrusions, or specifically de-
scribed as square-wave jerks, which are coupled with large microsaccades) [42]. More frequent
microsaccades during fixation, especially in the horizontal direction, are associated with a decreased
ability to hold fixation for long periods of time in PD [44]. Understanding microsaccade abnormalities in
PD is important because these types of eye movements are known to increase visual acuity [45] and
are critical for daily tasks such as reading [46]. Accordingly, microsaccade disturbances in PD might
be related to low-level visual functions such as contrast sensitivity [47] and performance in high-level
vision tasks such as visual search [48].

In contrast to considerable impairments in saccades, PD patients show only mild smooth

pursuit deficits, typically seen as decreased eye velocity relative to the target velocity, which
can then invoke saccades to allow the eyes to catch up to the target [8,49]. Not many studies

Trends in Neurosciences, January 2024, Vol. 47, No. 1 75

 Trends in Neurosciences

have examined the coordination of pursuit and saccades in PD. Some found comparatively
normal catch-up saccade rates during visually guided pursuit [47,50]. Others reported increased
catch-up saccade rates in PD patients and attributed these to a failure to inhibit saccades [51].
Whether changes in catch-up saccades during pursuit in patients are triggered by a retinal error
signal or whether they are a form of saccadic intrusion could be addressed by studies that include
a larger battery of different eye movement tests, further discussed in the following sections.

In general, pursuit deficits occur in healthy aging as well, and are therefore not necessarily specific
to PD, thus rendering visually guided pursuit of limited diagnostic value for PD. Impaired pursuit
appears to be more prevalent when tasks involve higher-level cognitive processing or deliberation,
such as remembering the motion path of a target (memory-based pursuit [11]). When patients with
preserved memory function had to remember the meaning of two consecutive visual cues, they
tended to use saccades rather than smooth pursuit to track the target [52]. Similarly, PD patients
cannot easily initiate smooth pursuit in cognitive anticipation of a visual target. This is shown in
studies that report fewer, slower, and delayed anticipatory eye movements in patients as
compared to healthy controls in tasks that included a target direction reversal [53] or required an-
ticipating the start of target motion [54]. Because the same function is relatively preserved in healthy
older people [55], this deficit appears to be disease-specific. Conversely, tasks that require
predicting the velocity of a previously shown target that is only briefly blanked from view yielded
only small performance differences between patients and controls [54]. Notably, even though pa-
tients tended to reach a lower peak velocity in these testing protocols, their eyes decelerated less
after the target had disappeared (i.e., maintained target velocity at a higher rate). These findings
show that patients can build up and maintain an internal representation of an object’s motion direc-
tion and velocity if they have been recently exposed to the same object (i.e., in the same trial).

Indeed, some oculomotor functions are preserved in PD patients. Most notably, this includes the
latency of visually guided saccades [56] and the initiation of visually driven smooth pursuit [52]. PD
patients also perform corrective saccades at a comparable level to healthy controls when they
must rapidly adapt to a new movement goal [57]. Congruently, movement adaptation is pre-
served when the trajectory of a hand movement must be corrected online [58]. All these tasks
have in common that they require a sense of urgency – a movement needs to be initiated or
corrected rapidly. When PD patients participated in a rapid go/no-go task that simulated the
real-world requirements of batting a baseball, their ball interception performance was comparable
to that of healthy controls, despite mild impairments in pursuit and saccadic tracking of the ball
[50]. Notably, the same patients exhibited expected performance impairments in saccade and
antisaccade tasks, including inhibitory dysfunction manifested in an abundance of premature
saccades to the wrong target (Figure 2).

We speculate that this differential preservation of function for rapid action in dynamic tasks might
imply the involvement of pathways that bypass dopaminergic connections through the basal gan-
glia, or a hyperdirect pathway linking frontal cortical eye movement areas to the basal ganglia
through the STN [59] (Figure 1). Although there is no direct evidence yet that links this pathway
to smooth pursuit, the STN is involved in pursuit and saccadic eye movement control and possi-
bly in impulse control function and reward-modulated goal pursuit [60]. Moreover, STN activity
serves switching from automatic to voluntary saccades to adapt to changes in task and situa-
tional demands [61]. It is directly implicated in the pathophysiology of PD [62] and serves as
the main target area for DBS [63,64], as we discuss in the following section.

We have focused so far in this section on conjugate eye movements – cases in which movements
of the eyes are yoked and go in the same direction – triggered by stimuli presented in the

76 Trends in Neurosciences, January 2024, Vol. 47, No. 1

Trends in Neurosciences

frontoparallel plane. In the real world, these movements are always combined with disconjugate
eye movements, such as vergence, to achieve binocular vision and to focus on objects of
interest across different depth planes. Vergence abnormalities result in impaired fusion and
double vision (Box 1), problems observed commonly in PD patients [65,66]. Congruently, vergence
eye movements are frequently impaired in these patients as compared to healthy age-matched
controls [66–69]. Reduced gain and increased convergence latency may be more pronounced
for vergence driven by disparity than for vergence driven by visual blur (accommodation) [66].
This differential impairment could help understand the effect of PD on the vergence neural path-
ways and mechanisms. The control of disparity- and blur-driven vergence has been assumed to
be partially separate anatomically, potentially relying on different brainstem nuclei and downstream
connections along a cerebro-ponto-cerebellar pathway [66,70], although this notion remains de-
bated [71]. The effect of subthalamic DBS on vergence in PD has the potential to further clarify
the neural substrate underlying vergence eye movements. Another interesting aspect about
vergence in PD is that patients compensate for the convergence deficit by using
different types of eye movements strategically [72]. As further evidence of plasticity in this sys-
tem, a preliminary study with a small group of PD patients reported recovery from convergence
insufficiency following vergence training [73], although a clinical trial would be required to con-
firm these results.

In summary, whereas PD patients commonly show significant impairments in saccade accuracy,

in the latency and error rate of antisaccades, and in vergence eye movements, pursuit can be
preserved and is therefore less often targeted in studies seeking to develop eye movement-
based biomarkers. This relative preservation of pursuit in conjunction with saccade impairments,
as well as differential impairments in vergence, can provide valuable insights into which specific
neuronal populations are impacted by the disease.

Eye movements used for diagnosis

Classifying and diagnosing movement disorders can be difficult due to their complex symptoms
and varied early clinical presentation. Yet, a correct classification lays the foundation for the subse-
quent diagnostic process, has prognostic value, and impacts treatment choices [3]. A clinical
diagnosis of PD requires not only the manifestation of cardinal symptoms – bradykinesia combined
with at least rest tremor or rigidity – but also the absence of exclusion criteria [5]. In this context, the
presence of specific eye movement abnormalities can assist in ruling out PD. Cerebellar oculomotor
abnormalities – sustained gaze-evoked nystagmus, macro square-wave jerks, and saccades
that are hypermetric (overshoot) – feature strongly in this category. These are abnormalities that
are commonly caused by cerebellar and brainstem disorders but are absent in PD [19]. Moreover,
slow or absent vertical saccades (vertical gaze palsy) in the downward direction rule out a diagnosis
of PD. By contrast, vertical gaze palsy combined with large square-wave jerks [41,42] can be a
characteristic symptom of PSP. In many instances, the diagnosis of PSP can be missed in the
early stages. Patients are often classified as having PD, because early PSP symptoms tend to
resemble those of idiopathic PD despite PSP being a tauopathy. In fact, slow vertical saccades
may be one of the first signs of PSP and are a hallmark symptom in many patients [74,75].
Combined with other symptoms, supranuclear vertical gaze palsy can also indicate dementia
with Lewy bodies, a Parkinsonian disorder that commonly causes dementia. Similarly, gaze-
evoked nystagmus can be present in other disorders within the parkinsonian spectrum such as, for
instance, multiple system atrophy. Abnormal nystagmus is diagnostically specific for this disorder,
because nystagmus is not usually present in either PD or PSP [19].

A systematic comparison between groups of patients with PD, PSP, and multiple system atrophy
as well as patients with rapid eye movement (REM) sleep behavior disorder – associated with the

Trends in Neurosciences, January 2024, Vol. 47, No. 1 77

 Trends in Neurosciences

prodromal stage of some parkinsonian disorders – revealed that saccade abnormalities manifest
early during the disease and differentiate PSP from other patient groups [15]. In this study, par-
ticipants engaged in simple, uninstructed free viewing of short video clips, a task that allows
measuring spontaneous and continuous eye movement behavior, including changes in pupil
size (Box 3). The gaze of patients with PSP was more biased toward the screen center than
that of other patient groups, and PSP patients exhibited fewer and smaller vertical eye movements
(saccades and microsaccades) than all other groups. Across all patients, vertical saccade
frequency and amplitude deficits correlated with motor symptom severity. The observed significant
reduction in vertical saccade rate and amplitude – pointing to involvement of midbrain regions
for vertical saccade control – differentiate PSP patients from other patient groups and suggest
differences in the pathophysiology between these disorders. Patients in the prodromal group
(REM sleep disorder group) revealed saccade alterations mimicking abnormalities observed in
the other patient groups, implying that saccade parameters could be used as early disease
markers [76]. To substantiate this conclusion, however, longitudinal studies are necessary, in
which patients are tracked as they convert from a prodromal stage to manifest disease, and this
process might take a considerable amount of time given the heterogeneity in disease progression.

A clinical bedside assessment of horizontal and vertical saccades, fixational stability and gaze-
evoked nystagmus may provide important information in the differential clinical diagnosis of PD.
However, the usefulness of eye movements in defining PD subtypes remains to be determined.
Subtyping relies entirely on clinical symptoms (87% of studies included in a systematic review;
[77]), and eye movements are not yet part of this assessment. Given the large variability in eye
movement abnormalities in PD patients, it may prove difficult to define parameter ranges that
separate one PD subtype from another. It may be that eye movements will contribute to a
biomarker-based classification approach of PD subtypes, in combination with other disease
indicators, rather than being used on their own.

Eye movements as tools to assess disease progression and treatment effects

The previous section focused on using eye movements as a potential tool to diagnose PD or to
differentiate PD from related disorders. Eye movement alterations might not only indicate disease
duration and motor burden [29], but can also reveal cognitive impairments, and thus help identify
patients who might be at risk for rapid cognitive decline [78]. Applying hierarchical cluster analysis
to data on latency and directional errors in the antisaccade task yielded different patterns of
cognitive abilities in PD patients. These different patient clusters, however, did not differ from
each other in terms of cognitive impairment. These findings indicate that various saccade perfor-
mance profiles might be related to similar cognitive function in PD patients [79] and provide an
important cautionary note that even though eye movement abnormalities may reliably occur in
PD, sensitivity and predictive power are currently lacking. However, in newly diagnosed patients,
eye movement tests can potentially provide insights into when cognitive functions begin to
decline [80] and help identify early signs of cognitive impairment, which in turn, serves as
a prognostic factor for disease severity. A study in newly diagnosed, drug-naive PD patients –
allowing the researchers to disentangle cognitive impairments related to disease from those
related to medication – revealed that antisaccades were already significantly impaired in early
PD, and in particular the antisaccade error rate was higher in patients than healthy controls
[81,82]. Some cognitive domains were also significantly impaired, even before medication was
initiated in this newly diagnosed patient group.

Moreover, eye movements could potentially be used to assess the effects of antiparkinsonian
medications or other treatments on motor control. The effect of medication on saccades has
been investigated by several groups. The most widely studied parameter is prosaccade latency,

78 Trends in Neurosciences, January 2024, Vol. 47, No. 1

Trends in Neurosciences

Box 3. Alterations in pupil physiology in PD

In a healthy person, the pupils dilate and constrict with changes in illumination. This pupil light reflex is the result of
antagonistic actions of the iris sphincter and dilator muscles, which are innervated, respectively, by the parasympathetic
and sympathetic nervous systems. In neurodiagnostic applications, the pupil light reflex is an important test of autonomic
nervous system function. As compared to age-matched, healthy controls, PD patients exhibit a delayed pupil light reflex
and a decrease in the amplitude, velocity, and acceleration of the constriction response [113]. Constriction velocity ap-
pears to be one of the most sensitive indicators in PD patients, linked to disease duration and severity [114]. Notably, al-
terations in the pupil light reflex occur in the absence of clinically noticeable autonomic nervous system dysfunctions,
indicating that pupil markers could potentially contribute to the diagnostic process during prodromal disease stages,
and before impairments in saccade parameters are seen [115].

Under constant light conditions, the pupil’s diameter also responds sensitively to a large range of cognitive processes
[116]. There is evidence that such cognitively modulated pupil responses are reduced in PD patients. In an interleaved
pro- and antisaccade task, pupil dilation in healthy controls depended on the type of task and response; that is, the pupil
was larger prior to stimulus appearance for correct antisaccades than for correct prosaccades or erroneous antisaccades.
PD patients’ pupil diameter depended less on task and response type, suggesting deficits in voluntary movement prepa-
ration in PD [117]. Similarly, PD patients show diminished changes in pupil size in response to different levels of expected
reward, and these effects are particularly pronounced when testing patients in an off-medication state [118].

Cognitive modulations of pupil size are served by a neural circuit receiving descending cortical inputs via the locus
coeruleus or the SC, which in turn receive cortical inputs from areas such as FEF [119]. This circuitry overlaps with some
of the areas and pathways impacted by PD, and thus, abnormalities in the pupil response in PD patients could be medi-
ated by loss of dopaminergic neurons in areas such as the locus coeruleus [120]. Pupil size is emerging as a useful tool to
index cognitive and autonomic function impairments in PD, although it is important to note that changes in the pupil light
reflex are not specific to PD. They also occur in other neurodegenerative diseases, such as Alzheimer’s disease [121], and
across a wide range of brain injuries, implying that they currently lack specificity to be useful as a biomarker to differentiate
between these disorders.

which is prolonged in PD and appears to be further prolonged when patients are on dopaminergic
medication compared to off medication [83–85]. The paradoxical increase in the degree of oculo-
motor abnormality, in the presence of a symptomatic treatment that can normalize somatic motor
function, is noteworthy. Of greater significance is the finding that antisaccade latency, while signifi-
cantly prolonged by PD, either improves [83] or is unaffected by symptomatic medication [85].
This may make it a useful biomarker of disease modifying drugs in future studies, and further work
is needed in this area.

Eye movements may also be affected by therapeutic electrical stimulation of parts of the basal ganglia
for treatment of motor symptoms of PD. In contrast to the effect of antiparkinsonian medication, DBS
of either the STN (Figure 1) or the globus pallidus internus (GPi) reduces prosaccade latency [86,87] in
parallel to improvements seen in motor symptoms. Whereas DBS improves accuracy in memory-
guided saccades [88], the effect of DBS on antisaccade error rate is less clear. Some studies suggest
that stimulation of the GPi, but not the STN, improves performance on the antisaccade task when
patients are on their antiparkinsonian medication [89]. By contrast, another study found that DBS
of the STN can impair antisaccade performance, and this was observed in patients who were off
medication when tested [90]. It is possible that these conflicting results stem from testing patients
in various states of medication as well as varied disease duration. Indeed, testing patients in ON
and OFF DBS states and ON and OFF dopaminergic medication yields three key observations
[87]. (i) DBS increases antisaccade error rate in patients who were OFF medication when tested. (ii)
DBS and medication state interact, and medication reduces the detrimental effects of DBS on
antisaccade performance. (iii) DBS effects on patients’ saccade task performance differ widely from
one patient to another, similarly to what has been described for antisaccade task performance, as
discussed in the preceding text (Figure 2). It appears that effects of DBS on saccade and antisaccade
performance are dependent on the individual (e.g., medication state and disease duration) and testing
context as well as potentially on the exact location [91] and frequency of stimulation [92]. In
addition, DBS of the STN disrupts the normal modulation of saccadic reaction time according

Trends in Neurosciences, January 2024, Vol. 47, No. 1 79

 Trends in Neurosciences

to target probability, in line with the notion that the basal ganglia perform Bayesian statistics for Outstanding questions
competing movement selection [93]. When asked to saccade to a target in one of two locations Researchers are beginning to
associated with a low or high probability of target appearance, saccade latencies to the likely understand the neural mechanisms that
underlie changes in eye movements in
target location were shorter than those to the unlikely location for healthy controls and
PD. How are these changes related to
patients OFF stimulation. However, ON stimulation, patients were no longer able to tailor disease progression, and how are they
their responses to stimulus probability, and saccade latencies to unlikely stimuli did not impacted by medication or DBS?
increase as they were expected to [94].
Whereas saccades are commonly
impaired in PD, other oculomotor func-
Even though the scientific and clinical focus has been on the assessment of saccades and tions, such as smooth pursuit, can be
antisaccades, studies have also looked at the effect of STN DBS on fixational eye movements preserved. Are there specific neuronal
in PD [95] and found that the amplitude and frequency of microsaccades are increased following populations in areas that jointly control
saccades and pursuit that are impacted
DBS. In many patients, DBS also increased the prevalence of square-wave jerks, but with high by PD?
interindividual variability in results. Moreover, STN DBS might alter the temporal pattern of
microsaccades. Microsaccades in general are temporally more variable (i.e., less regular or Deficits in visual function (e.g., acuity,
clustered) in PD than in healthy controls. DBS can partially restore microsaccade rhythmicity contrast sensitivity) can occur as non-
motor symptoms in PD, and are
and clusteredness in PD patients [43]. These results demonstrate that the STN may play a role associated with dopamine deficiency
in modulating the temporal dynamics of microsaccades. and related morphological changes in
the retina. Assessing visual and oculo-
motor abnormalities within the same
Few studies have looked at the effects of DBS on smooth pursuit in PD or related disorders.
studies and tasks can help characterize
Assessing oculomotor control in a group of nine PD patients who had undergone STN DBS underlying mechanisms and address
surgery and were tested OFF medication found reduced smooth pursuit gain with DBS off, the potential link between the two. Are
whereas switching the stimulator on brought the gain back up toward unity [96]. Likewise, pursuit visual abnormalities related to or even
caused by oculomotor abnormalities?
velocity accuracy, defined as the percentage of the time during which pursuit velocity was within
20% of target velocity, was increased by DBS. These results demonstrate the potential of DBS to To ensure consistent results across
mitigate pursuit dysfunction in PD. Given the therapeutic effects that DBS of the basal ganglia has studies, eye movement measurement,
on saccades and pursuit in PD, it will be important to investigate in the future whether these interpretation, and reporting need to
follow standard protocols. Reaching
benefits extend to vergence eye movements as well. methodological consensus is easier
for simplified tasks such as pro- or
In summary, the modulatory effects of subthalamic DBS on the characteristics of saccades, antisaccades, but can be more chal-
microsaccades, and pursuit serve not only as important treatment indicators but also provide a lenging for complex paradigms. How
can eye movement protocols that in-
unique opportunity to investigate how the basal ganglia interact with other neural structures corporate naturalistic task requirements
that control eye movements [97,98]. This knowledge, together with our understanding of the be standardized?
neurophysiology of eye movements, can help elucidate the mechanism of action of DBS, which
The sensitivity and specificity of eye
are still not well understood.
movements as a biomarker of PD
need to be further determined. How
Concluding remarks well can eye movements help distin-
PD is a heterogeneous disorder where diagnosis, progression, and treatment are currently moni- guish PD from other conditions with
shared symptoms?
tored using standardized clinical rating scales such as the Movement Disorder Society-Unified
Parkinson’s Disease Rating Scale (MDS-UPDRS [99]). Although clinical examination of saccades Eye movements are beginning to be
has, for a long time, formed part of the assessment of neurodegenerative conditions such as used to quantify neurodegenerative
Huntington’s disease and PSP, it does not feature in scales for assessing PD despite the evidence disease and monitor its progression.
Can eye movements be used to
that eye movements are abnormal [9,10]. Future translational research in PD will be dependent on measure the efficacy of treatment in
finding more precise and objective measures of disease that follow standardized protocols (see clinical trials?
Outstanding questions). Measurements of eye movement abnormalities are strong candidates
for this. Investigators are increasingly turning to more complex multivariate approaches and
machine learning algorithms to address reliability concerns [28,100]. Multicenter and large
collaborative studies [16,32] including thousands of patients and consensus-based eye movement
protocols pave the way to big data approaches, linking clinical, behavioral, and cellular-molecular
dimensions. Importantly, differential impairments in saccades and pursuit eye movements in PD
allow testing assumptions about different populations of neurons along overlapping brain

80 Trends in Neurosciences, January 2024, Vol. 47, No. 1

Trends in Neurosciences

pathways that may be impacted by the disease. Treatment approaches such as DBS, in addition
to their therapeutic value, provide scientists with a unique opportunity to systematically probe and
investigate the brain networks that underlie human eye movement control.

Acknowledgments
The authors thank Silke Cresswell, James FitzGerald, and members of the UBC Oculomotor Laboratory for helpful comments on
the manuscript. C.A.A. was supported by the National Institute for Health and Care Research Oxford Biomedical Research Centre.
M.S.’s foundational research is funded by a Discovery Grant of the Natural Sciences and Engineering Research Council of Canada.

Declaration of interests
The authors have no conflicts of interest to declare.

References
1. Dorsey, E.R. et al. (2018) Global, regional, and national bur- 22. Hikosaka, O. et al. (2019) Direct and indirect pathways for
den of Parkinson’s disease, 1990–2016: a systematic anal- choosing objects and actions. Eur. J. Neurosci. 49, 637–645
ysis for the Global Burden of Disease Study 2016. Lancet 23. Yasuda, M. and Hikosaka, O. (2017) To wait or not to wait-
Neurol. 17, 939–953 separate mechanisms in the oculomotor circuit of basal ganglia.
2. Tolosa, E. et al. (2006) The diagnosis of Parkinson’s disease. Front. Neuroanat. 11, 35
Lancet Neurol. 5, 75–86 24. Stuphorn, V. and Schall, J.D. (2006) Executive control of
3. Abdo, W.F. et al. (2010) The clinical approach to movement countermanding saccades by the supplementary eye field.
disorders. Nat. Rev. Neurol. 6, 29–37 Nat. Neurosci. 9, 925–931
4. Marsili, L. et al. (2018) Diagnostic criteria for Parkinson’s dis- 25. Sajad, A. et al. (2019) Cortical microcircuitry of performance
ease: from James Parkinson to the concept of prodromal dis- monitoring. Nat. Neurosci. 22, 265–274
ease. Front. Neurol. 9, 156 26. Jahanshahi, M. and Rothwell, J.C. (2017) Inhibitory dysfunction
5. Postuma, R.B. et al. (2015) MDS clinical diagnostic criteria for contributes to some of the motor and non-motor symptoms of
Parkinson’s disease. Mov. Disord. 30, 1591–1601 movement disorders and psychiatric disorders. Philos. Trans.
6. Fereshtehnejad, S.M. et al. (2017) Clinical criteria for subtyping R. Soc. Lond. B 372, 1718
Parkinson’s disease: biomarkers and longitudinal progression. 27. Verbruggen, F. et al. (2019) A consensus guide to capturing the
Brain 140, 1959–1976 ability to inhibit actions and impulsive behaviors in the stop-
7. Han, S.C. and Frucht, S.J. (2022) Atypical Parkinsonian syn- signal task. eLife 8, e46323
dromes: awareness of red flags and key examination findings 28. Bredemeyer, O. et al. (2022) Oculomotor deficits in Parkinson’s
speeds diagnosis. Pract. Neurol. 7, 44–48 disease: increasing sensitivity using multivariate approaches.
8. Pinkhardt, E.H. and Kassubek, J. (2011) Ocular motor abnormalities Front. Digit. Health 4, 939677
in Parkinsonian syndromes. Parkinsonism Relat. Disord. 17, 29. Waldthaler, J. et al. (2021) Antisaccades in Parkinson’s disease:
223–230 a meta-analysis. Neuropsychol. Rev. 31, 628–642
9. Armstrong, R.A. (2015) Oculo-visual dysfunction in Parkinson’s 30. Antoniades, C.A. et al. (2013) An internationally standardised
disease. J. Parkinsons Dis. 5, 715–726 antisaccade protocol. Vis. Res. 84, 1–5
10. Kassavetis, P. et al. (2022) Eye movement disorders in move- 31. Degos, B. et al. (2022) From anticipation to impulsivity in
ment disorders. Mov. Disord. Clin. Pract. 9, 284–295 Parkinson’s disease. NPJ Parkinsons Dis. 8, 125
11. Fukushima, K. et al. (2017) Clinical application of eye movement 32. Riek, H.C. et al. (2023) Cognitive correlates of antisaccade
tasks as an aid to understanding Parkinson’s disease patho- behaviour across multiple neurodegenerative diseases. Brain
physiology. Exp. Brain Res. 235, 1309–1321 Commun. 5, fcad049
12. Terao, Y. et al. (2017) What do eye movements tell us about patients 33. Crawford, T.J. et al. (1989) Abnormalities of nonvisually-guided
with neurological disorders? — an introduction to saccade recording eye movements in Parkinson’s disease. Brain 112, 1573–1586
in the clinical setting. Proc. Jpn. Acad. Ser. B 93, 772–801 34. Lueck, C.J. et al. (1992) Saccadic eye movements in
13. Mertin, R. et al. (2023) Oculomotor abnormalities indicate early ex- Parkinson’s disease: II. Remembered saccades–towards a uni-
ecutive dysfunction in prodromal X-linked dystonia-parkinsonism fied hypothesis? Q. J. Exp. Psychol. 45, 211–233
(XDP). J. Neurol. 270, 4262–4275 35. Ketcham, C.J. et al. (2003) Memory-motor transformations are
14. FitzGerald, J.J. et al. (2018) Quantifying motor impairment in impaired in Parkinson’s disease. Exp. Brain Res. 149, 30–39
movement disorders. Front. Neurosci. 12, 202 36. Vermersch, A.I. et al. (1994) Sequences of memory-guided
15. Habibi, M. et al. (2022) Eye tracking identifies biomarkers in α- saccades in Parkinson’s disease. Ann. Neurol. 35, 487–490
synucleinopathies versus progressive supranuclear palsy. 37. Middleton, F.A. and Strick, P.L. (2000) Basal ganglia and
J. Neurol. 269, 4920–4938 cerebellar loops: motor and cognitive circuits. Brain Res. Rev. 31,
16. Brien, D.C. et al. (2023) Classification and staging of Parkinson’s 236–250
disease using video-based eye tracking. Parkinsonism Relat. 38. Martinez-Conde, S. et al. (2013) The impact of microsaccades
Disord. 110, 105316 on vision: towards a unified theory of saccadic function. Nat.
17. Land, M. (2019) Eye movements in man and other animals. Vis. Rev. Neurosci. 14, 83–96
Res. 162, 1–7 39. Otero-Millan, J. et al. (2013) An oculomotor continuum from explora-
18. Spering, M. and Carrasco, M. (2015) Acting without seeing: eye tion to fixation. Proc. Natl. Acad. Sci. U. S. A. 110, 6175–6180
movements reveal visual processing without awareness. 40. Krauzlis, R.J. et al. (2017) Neuronal control of fixation and fixational
Trends Neurosci. 38, 247–258 eye movements. Philos. Trans. R. Soc. B 372, 20160205
19. Leigh, R.J. and Zee, D.S. (2015) The Neurology of Eye Move- 41. Otero-Millan, J. et al. (2013) Saccades during attempted fixation in
ments (5th ed.), Oxford University Press parkinsonian disorders and recessive ataxia: from microsaccades
20. Le Heron, C. et al. (2018) The anatomy of apathy: a neurocog- to square-wave jerks. PLoS One 8, e58535
nitive framework for amotivated behaviour. Neuropsychologia 42. Alexander, R.G. et al. (2018) Microsaccade characteristics in
118, 54–67 neurological and ophthalmic disease. Front. Neurol. 9, 144
21. Muhammed, K. et al. (2021) Reward sensitivity and action in 43. Beylergil, S.B. et al. (2022) Temporal patterns of spontaneous
Parkinson’s disease patients with and without apathy. Brain fixational eye movements: the influence of basal ganglia.
Commun. 3, fcab022 J. Neuroophthalmol. 42, 45–55

Trends in Neurosciences, January 2024, Vol. 47, No. 1 81

 Trends in Neurosciences

44. Tsitsi, P. et al. (2021) Fixation duration and pupil size as 73. Machan, C.M. et al. (2020) Changing vergence function in per-
diagnostic tools in Parkinson’s disease. J. Parkinsons Dis. 11, sons with Parkinson's disease and convergence insufficiency.
865–875 Parkinsonism Relat. Disord. 73, 41–43
45. Intoy, J. and Rucci, M. (2020) Finely tuned eye movements 74. Chen, A.L. et al. (2010) The disturbance of gaze in progressive
enhance visual acuity. Nat. Commun. 11, 795 supranuclear palsy: implications for pathogenesis. Front.
46. Bowers, N.R. and Poletti, M. (2017) Microsaccades during Neurol. 1, 147
reading. PLoS One 12, e0185180 75. Terao, Y. et al. (2021) Deciphering the saccade velocity pro-
47. Ming, W. et al. (2016) Visual contrast sensitivity in early- file of progressive supranuclear palsy: a sign of latent cere-
stage Parkinson’s disease. Invest. Ophthalmol. Vis. Sci. bellar/brainstem dysfunction? Clin. Neurophysiol. 21,
57, 5696–5704 37–47
48. Beylergil, S.B. et al. (2022) Eye movements in Parkinson’s 76. Heller, J. et al. (2017) Brain imaging findings in idiopathic REM
disease during visual search. J. Neurol. Sci. 440, 120299 sleep behavior disorder (RBD) - a systematic review on potential
49. de Brouwer, S. et al. (2002) What triggers catch-up saccades biomarkers for neurodegeneration. Sleep Med. Rev. 34, 23–33
during visual tracking? J. Neurophysiol. 87, 1646–1650 77. Mestre, T.A. et al. (2021) Parkinson’s disease subtypes: critical
50. Fooken, J. et al. (2022) Preservation of eye movements in appraisal and recommendations. J. Parkinsons Dis. 11,
Parkinson's disease are stimulus and task specific. J. Neurosci. 395–404
42, 487–499 78. Stuart, S. et al. (2019) Pro-saccades predict cognitive decline in
51. Wu, C.C. et al. (2018) Eye movement control during visual Parkinson’s disease: ICICLE-PD. Mov. Disord. 34, 1690–1698
pursuit in Parkinson’s disease. PeerJ 6, e5442 79. Waldthaler, J. et al. (2023) Cluster analysis reveals distinct
52. Fukushima, K. et al. (2015) Impaired smooth-pursuit in patterns of saccade impairment and their relation to cognitive
Parkinson’s disease: normal cue-information memory, but profiles in Parkinson’s disease. J. Neuropsychol. 17, 251–263
dysfunction of extra-retinal mechanisms for pursuit preparation 80. Ouerfelli-Ethier, J. et al. (2018) Anti-saccades predict cognitive
and execution. Physiol. Rep. 3, e12361 functions in older adults and patients with Parkinson’s disease.
53. de Hemptinne, C. et al. (2013) How does Parkinson’s disease PLoS One 13, e0207589
and aging affect temporal expectation and the implicit timing 81. Antoniades, C.A. et al. (2015) Antisaccades and executive dys-
of eye movements? Neuropsychologia 51, 340–348 function in early drug-naive Parkinson’s disease: the discovery
54. Helmchen, C. et al. (2012) Role of anticipation and prediction in study. Mov. Disord. 30, 843–847
smooth pursuit eye movement control in Parkinson's disease. 82. Zhou, M.X. et al. (2022) Oculomotor impairments in de novo
Mov. Disord. 27, 1012–1018 Parkinson’s disease. Front. Aging Neurosci. 14, 985679
55. Sprenger, A. et al. (2011) The role of prediction and anticipation 83. Hood, A.J. et al. (2007) Levodopa slows prosaccades and
on age-related effects on smooth pursuit eye movements. Ann. improves antisaccades: an eye movement study in Parkinson’s
N. Y. Acad. Sci. 1233, 168–176 disease. J. Neurol. Neurosurg. Psychiatry 78, 565–570
56. Chan, F. et al. (2005) Deficits in saccadic eye-movement control 84. Lu, Z. et al. (2019) The effect of levodopa on saccades - Oxford
in Parkinson’s disease. Neuropsychologia 43, 784–796 Quantification in Parkinsonism study. Parkinsonism Relat. Disord.
57. Merritt, K.E. et al. (2017) Automatic online motor control is intact 68, 49–56
in Parkinson’s disease with and without perceptual awareness. 85. Munoz, M.J. et al. (2022) Medication adversely impacts visually-
eNeuro 4 ENEURO.0215-17.2017-12 guided eye movements in Parkinson’s disease. Clin. Neurophysiol.
58. Desmurget, M. et al. (2004) On-line motor control in patients 143, 145–153
with Parkinson’s disease. Brain 127, 1755–1773 86. Temel, Y. et al. (2008) Saccadic latency during electrical stimulation
59. Chen, W. et al. (2020) Prefrontal-subthalamic hyperdirect pathway of the human subthalamic nucleus. Curr. Biol. 18, R412–R414
modulates movement inhibition in humans. Neuron 106, 579–588 87. Bakhtiari, S. et al. (2020) The role of the subthalamic nucleus in
60. Wessel, J.R. et al. (2022) A causal role for the human subtha- inhibitory control of oculomotor behavior in Parkinson’s dis-
lamic nucleus in non-selective cortico-motor inhibition. Curr. ease. Sci. Rep. 10, 5429
Biol. 32, 3785–3791 88. Rivaud-Péchoux, S. et al. (2000) Improvement of memory guided
61. Isoda, M. and Hikosaka, O. (2008) Role for subthalamic nucleus saccades in parkinsonian patients by high frequency subthalamic
neurons in switching from automatic to controlled eye move- nucleus stimulation. J. Neurol. Neurosurg. Psychiatry 68, 381–384
ment. J. Neurosci. 28, 7209–7218 89. Antoniades, C.A. et al. (2015) Pallidal deep brain stimulation im-
62. Oswal, A. et al. (2021) Neural signatures of hyperdirect pathway proves higher control of the oculomotor system in Parkinson’s
activity in Parkinson’s disease. Nat. Commun. 12, 5185 disease. J. Neurosci. 35, 13043–13052
63. Horn, A. et al. (2019) Deep brain stimulation induced normalization 90. Goelz, L.C. et al. (2017) The effects of unilateral versus bilateral
of the human functional connectome in Parkinson’s disease. subthalamic nucleus deep brain stimulation on prosaccades
Brain 142, 3129–3143 and antisaccades in Parkinson’s disease. Exp. Brain Res.
64. Krauss, J.K. et al. (2021) Technology of deep brain stimulation: 235, 615–626
current status and future directions. Nat. Rev. Neurol. 17, 91. Waldthaler, J. et al. (2023) Stimulation of non-motor subtha-
75–87 lamic nucleus impairs selective response inhibition via prefrontal
65. Irving, E.L. et al. (2016) Prevalence of convergence insufficiency connectivity. Brain Commun. 5, fcad121
in Parkinson’s disease. Mov. Disord. Clin. Pract. 4, 424–429 92. Zacharia, A. et al. (2023) Effects of deep brain stimulation
66. Gupta, P. et al. (2021) Effects of Parkinson disease on blur- frequency on eye movements and cognitive control. NPJ
driven and disparity-driven vergence eye movements. Parkinsons Dis. 9, 50
J. Neuroophthalmol. 41, 442–451 93. Mink, J.W. (1996) The basal ganglia: focused selection and
67. Hanuska, J. et al. (2015) Fast vergence eye movements are inhibition of competing motor programs. Prog. Neurobiol. 50,
disrupted in Parkinson’s disease: a video-oculography study. 381–425
Parkinsonism Relat. Disord. 21, 797–799 94. Antoniades, C.A. et al. (2014) Deep brain stimulation
68. Kang, S.L. et al. (2018) Vergence and strabismus in neuro- abolishes slowing of reactions to unlikely stimuli. J. Neurosci.
degenerative disorders. Front. Neurol. 9, 299 34, 10844–10852
69. Sun, Y.R. et al. (2023) Monitoring eye movement in patients with 95. Beylergil, S.B. et al. (2021) Effects of subthalamic deep brain
Parkinson’s disease: what can it tell us? Eye Brain 15, 101–112 stimulation on fixational eye movements in Parkinson’s disease.
70. Gamlin, P.D. et al. (1996) The role of cerebro-ponto-cerebellar path- J. Comput. Neurosci. 49, 345–356
ways in the control of vergence eye movements. Eye 10, 167–171 96. Nilsson, M.H. et al. (2013) Subthalamic deep brain stimulation
71. Cullen, K.E. and Van Horn, M.R. (2011) The neural control of improves smooth pursuit and saccade performance in patients
fast vs. slow vergence eye movements. Eur. J. Neurosci. 33, with Parkinson’s disease. J. Neuroeng. Rehabil. 10, 33
2147–2154 97. FitzGerald, J.J. and Antoniades, C.A. (2016) Eye movements
72. Gupta, P. et al. (2021) Computational models to delineate 3D and deep brain stimulation. Curr. Opin. Neurol. 29, 69–73
gaze-shift strategies in Parkinson’s disease. J. Neural Eng. 18, 98. Shaikh, A.G. et al. (2018) Effects of deep brain stimulation on
0460a5 eye movements and vestibular function. Front. Neurol. 9, 444

82 Trends in Neurosciences, January 2024, Vol. 47, No. 1

Trends in Neurosciences

99. Goetz, C.G. et al. (2007) Movement Disorder Society-spon- 110. Hikosaka, O. and Wurtz, R.H. (1983) Visual and oculomotor
sored revision of the Unified Parkinson’s Disease Rating Scale functions of monkey substantia nigra pars reticulata. IV. Rela-
(MDS-UPDRS): process, format, and clinimetric testing plan. tion of substantia nigra to superior colliculus. J. Neurophysiol.
Mov. Disord. 22, 41–47 49, 1285–1301
100. Przybyszewski, A.W. et al. (2023) Machine learning and eye 111. Hikosaka, O. et al. (2000) Role of the basal ganglia in the control of
movements give insights into neurodegenerative disease purposive saccadic eye movements. Physiol. Rev. 80, 953–978
mechanisms. Sensors 23, 2145 112. Ryczko, D. and Dubuc, R. (2023) Dopamine control of down-
101. Shaikh, A.G. and Ghasia, F.F. (2019) Saccades in Parkinson’s dis- stream motor centers. Curr. Opin. Neurobiol. 83, 102785
ease: hypometric, slow, and maladaptive. Prog. Brain Res. 249, 113. Micieli, G. et al. (1991) Disordered pupil reactivity in Parkinson’s
81–94 disease. Clin. Auton. Res. 1, 55–58
102. Amador, S.C. et al. (2006) Dissociating cognitive deficits in- 114. You, S. et al. (2021) Analysis of pupillometer results according
volved in voluntary eye movement dysfunctions in Parkinson’s to disease stage in patients with Parkinson’s disease. Sci.
disease patients. Neuropsychologia 44, 1475–1482 Rep. 11, 17880
103. Lemos, J. and Eggenberger, E. (2013) Saccadic intrusions: 115. Perkins, J.E. et al. (2021) Saccade, pupil, and blink responses
review and update. Curr. Opin. Neurol. 26, 59–66 in rapid eye movement sleep behavior Disorder. Mov. Disord.
104. Kaski, D. and Bronstein, A.M. (2017) Ocular tremor in 36, 1720–1726
Parkinson’s disease: discussion, debate, and controversy. 116. Ebitz, R.B. and Moore, T. (2019) Both a gauge and a filter: cog-
Front. Neurol. 8, 134 nitive modulations of pupil size. Front. Neurol. 9, 1190
105. Kowler, E. (2011) Eye movements: the past 25 years. Vis. Res. 117. Wang, C.A. et al. (2016) Disruption of pupil size modu-
51, 1457–1483 lation correlates with voluntary motor preparation defi-
106. Spering, M. and Montagnini, A. (2011) Do we track what we cits in Parkinson’s disease. Neuropsychologia 80,
see? Common versus independent processing for motion 176–184
perception and smooth pursuit eye movements: a review. Vis. 118. Manohar, S.G. and Husain, M. (2015) Reduced pupillary
Res. 51, 836–852 reward sensitivity in Parkinson’s disease. NPJ Parkinsons
107. Munoz, D.P. and Everling, S. (2004) Look away: the anti- Dis. 1, 15026
saccade task and the voluntary control of eye movement. 119. Lehmann, S.J. and Corneil, B.D. (2016) Transient pupil dilation
Nat. Rev. Neurosci. 5, 218–228 after subsaccadic microstimulation of primate frontal eye fields.
108. Everling, S. et al. (1999) Role of primate superior colliculus in J. Neurosci. 36, 3765–3776
preparation and execution of anti-saccades and pro- 120. Sun, J. et al. (2023) Disruption of locus coeruleus-related func-
saccades. J. Neurosci. 19, 2740–2754 tional networks in Parkinson’s disease. NPJ Parkinsons Dis. 9, 81
109. Everling, S. and Munoz, D.P. (2000) Neuronal correlates for pre- 121. Bittner, D.M. et al. (2014) Repetitive pupil light reflex: potential
paratory set associated with pro-saccades and anti-saccades marker in Alzheimer’s disease? J. Alzheimer Dis. 42,
in the primate frontal eye field. J. Neurosci. 20, 387–400 1469–1477

Trends in Neurosciences, January 2024, Vol. 47, No. 1 83