menu Contact Us Dizzy Patients Health Care Providers Dizzy Culture Research Site DVD BPPV DVD Tai Chi DVD Understanding Dizziness Acknowledgements Disclaimer Quoting

TRACKING TEST (Smooth Pursuit)

Timothy C. Hain, MD

Page last modified: July 1, 2012

The screen shot above, from our clinical laboratory in Chicago, shows what a computerized tracking test looks like. The tracking test measures the ability of subjects to match eye movement to visual target movement. The blue line is a target projected by an LCD projector on as screen about 1.5 meters in front of the subject. The green line on the top is the horizontal eye position. Note that sometimes the eye falls behind the target and has to make an abrupt more rapid movement to catch up (this is called a "catch-up saccade). The vertical eye position is shown on the bttom trace. It is rather quiet -- showing occasional blink artifacts, and a small amount of undulation related an imprecise rotational alignment of the eye-camera with respect to the left eye (this can be easily seen on the right video, which contains the left eye).

A example of a processed tracking test (labelled Smooth Pursuit) is shown below. Sinusoidal pursuit is processed with Fourier fits and a Bode plot is produced.  This figure shows normal pursuit, where only one eye was recorded.


 Methods of producing a smooth pursuit target.


Pendulum that can be used for eliciting sinusoidal smooth pursuit. The target can be a brightly colored sphere, such as a golf ball. Laser projection device that can be used as a pursuit stimulus. This version is sold by Neurokinetics.. Other methods of producing pursuit stimuli include digital projection systems and light bars.

The figure above shows two examples of methods of producing a smooth pursuit target. The "low tech" method is to attach a brightly colored heavy object, such as a tennis ball or golf ball, to a string. This can be used to produce a sinusoidally moving target, at a single frequency which is determined by the length of the string. The peak velocity is variable with this method. Only sinusoidal stimuli are possible.

A more reproducible method of producing a smooth pursuit target is to use a laser or a LCD projector to project a target on a screen. As the price of LCD projectors has come down dramatically in recent years, this is presently the preferred method.

Light bar. An inferior method of eliciting smooth pursuit


An older method is to use an "LED bar". This is an array of LED's arranged in an arc. By illuminating them sequentially, the illusion of smooth movement can be created, and by turning individual LED's on, the saccade test can be performed. At this writing, we think that the LCD projector method is far superior.


Both the sinusoidal and the triangular wave pursuit stimuli are used for clinical testing. Sinusoidal stimuli are appropriate for detecting symmetrical disturbances of pursuit and triangular wave stimuli are used to detect pursuit which is better in one direction than the other. Pursuit gain, which is the ratio of eye velocity to target velocity, is affected by target velocity, acceleration and frequency. For the sinusoidal pursuit stimulus, these three stimulus parameters are mutually interdependent. For the triangular wave pursuit stimulus, velocity is constant, and acceleration appears as periodic pulses. Accordingly, frequency and velocity can be varied independently of acceleration. Unfortunately, perfect tracking of the triangular wave stimulus is impossible because of the abrupt accelerations at turn-around time.

Registration of smooth pursuit is of minor diagnostic utility, because disturbances of pursuit are usually nonspecific. Pursuit performance is strongly affected by attention, and inattentive or uncooperative subjects can perform poorly without having any significant central lesion. Another source of difficulty is that the lack of a standard pursuit paradigm associated with a well defined normal data set. Simple sinusoidal pursuit paradigms can be characterized by pair of three variables (frequency, amplitude and peak velocity), and pursuit tracking performance is a function of all three variables. Most laboratories have used idiosyncratic combinations of paradigm variables, which has resulted in the generation of many small normal data sets which cannot be compared to others. There is considerable variability even when the paradigm variables are similar. This variability may be related to factors that are difficult to quantify such as the degree of alertness present in subjects, or the visibility of the pursuit target. Pursuit is easily disrupted by common centrally acting medications such as anticonvulsants, minor tranquilizers, and preparations used for sleep. Finally, it is also clear that pursuit performance declines with age (Zackon and Sharpe, 1987).


Normal Limits for Smooth Pursuit

Sinuosoidal smooth pursuit is strongly reduced by frequency, age and gender (women have poorer pursuit than men).

Eyeball method of pursuit analysis

Different brands of clinical EOG systems succeed to a lesser or greater extent in quantifying smooth pursuit. Usually the error is on the side of normal -- persons with poor pursuit are scored as having better tracking than is correct. Algorithms to detect poor calibrations are also presently quite primitive. Because of these methodological problems, and just to be safe, it is always preferable to just "eye-ball" the pursuit trace, rather than to depend on the extremely variable performance of commercial systems.

The questions one should ask are:

More detail about how one figures out the answers to these questions is found below.

Calibration Errors

Calibration error is easily seen in this pursuit recording. The eye (red/green) is apparently tracking a target moving at roughly twice the excursion of the target (blue)



It is easily possible to use an incorrect calibration during ENG/VNG testing, and when this is done, it is easily seen on smooth pursuit. An example is shown in the figure above. When this is seen, the operator should recalibrate. This can sometimes be done after the recording is done, as pursuit is intrinsically "self calibrating".

Symmetrical Disturbances of Pursuit

Disorders of Smooth Pursuit

Symmetrical reduction of smooth pursuit (reduced gain) is encountered very frequently. The table above lists the most common causes of reduced pursuit gain. For the reasons advanced above, one should be conservative when diagnosing abnormalities of pursuit. Clinically, it is adequate to classify patients with symmetrical pursuit into those with perfect pursuit, those with moderately impaired pursuit, and those with no pursuit at all. This classification can usually be done by eye from the position trace, when one uses a reasonable sinusoidal stimulus (e.g. 0.5 Hz, +- 20 degree amplitude). Cerebellar lesions have significant but relatively minor effects on pursuit (Straube et al, 1997).

Those with perfect or near perfect pursuit, as judged from the lack of catch-up saccades, or from pursuit gains greater than 0.8, are normal. Persons with some, but not perfect pursuit, e.g. pursuit gains greater than 0.2 but less than 0.8 are in a grey zone. Such moderately impaired pursuit tracking might be related to inattention or medication, an underlying central nervous system disorder, or advanced age.

Persons with no pursuit at all, operationally defined as pursuit gain less than 0.2, are the most important ones to identify because they will nearly always have a central nervous system disturbance. Rarely, pursuit gain greater than 1.0 is noted. This is recognized by the occurrence of "backup" saccades, or in other words, saccades directed against target motion. If backup saccades are not present, one will inevitably find a technical error. Pursuit gain which is truly greater than 1.0 occurs most frequently in patients with a form of congenital nystagmus called "latent nystagmus", during triangular wave pursuit. Some normal subjects can also track with gains slightly greater than 1.0.

One must have central vision to pursue. Thus blind people cannot pursue targets, and people whose vision has been obscured by cataract or removal of their habitual spectacles. On the other hand, persons with good central vision, but bad peripheral vision may be able to pursue quite well. The following figure shows a patient who has terrible peripheral vision (see their recording under the saccadic disorder page), but has good central vision. For this reason, they can pursue fairly well.


Patients with retinitis pigmentosa (RP). In this condition peripheral vision is poor, but central vision may be good. They have "tunnel vision". This causes a peculiar situation where saccades may be very disorganized, but pursuit may be normal.

Reduced smooth pursuit -- drug ingestion.

While we know of no attempt to test every single medication for its effect on pursuit, as far as we know, any sedating medication can impair smooth pursuit. To be sedating a medication has to get into the brain. Stimulant medications generally do not impair smooth pursuit but cause extraneous saccades instead (square wave jerks)

Frequent offenders are medications for dizziness (e.g. meclizine, benzodiazepines), psychiatric medications, and neurology medications. An example of the effect of Olanzapine on smooth pursuit is shown below (image contributed by Dr. Dario Yacovino).

Olanzapine 0.4 hz

Asymmetrical Pursuit

Causes of Asymmetrical Pursuit


Pursuit which is significantly worse in one direction than another is asymmetrical. While rare, asymmetrical pursuit is more often of clinical utility than is symmetrically reduced pursuit, because it is specific for a central nervous system disorder. One can easily detect pursuit asymmetry if a plot is available in which there is an indication of mean gain and the standard deviation in each direction. One must use a pursuit stimulus in which velocity is constant, such as the triangular wave paradigm, in order to be able to compare rightward and leftward gain.

There are several causes of asymmetrical pursuit, as listed in the table aove. Patients with acute cortical lesions involving the parietal or frontal lobes, or a brainstem lesion involving the pontine nuclei may transiently exhibit better pursuit directed contralateral to their lesion. Pursuit asymmetry due to a cortical injury is rarely useful because it typically persists only for several weeks. After several months, there may still be inattention to the side opposite the lesion.

Unidirectional spontaneous nystagmus may be superimposed upon pursuit and cause asymmetry. Spontaneous nystagmus due to peripheral vestibular lesions, when weak, may not affect pursuit at all but when it is strong (e.g. 20 deg/sec in the dark), it may overwhelm the pursuit system. Spontaneous nystagmus due to central lesions may go uncorrected by the pursuit system, and result in a pronounced asymmetry pattern. These patients present with a spontaneous nystagmus, poorly suppressed by fixation, reduced and asymmetrical pursuit, and gaze-evoked nystagmus. In these instances it is helpful to measure pursuit gain only around regions where the eye is crossing primary position, as in this way, the effects of gaze can be eliminated.

In patient with a form of congenital nystagmus called latent nystagmus, a pursuit asymmetry can be recorded which alternates direction according to the viewing eye. These patients usually have a history of amblyopia. As no pursuit asymmetry or nystagmus may be seen with both eyes viewing, this condition can cause confusion if the patient alternates fixation during the oculomotor battery.

Reversed Pursuit

In certain patients with congenital nystagmus, the eyes will make saccades in the direction of target motion and slow smooth movements against target motion. Some authors prefer avoid using the term "reversed pursuit" because eye velocity is not proportional to target velocity in these patients (Abadie and Dickenson, 1981).

Miscellaneous Pursuit Abnormalities.

Disorganized vertical pursuit.. Although this patient can clearly track with reasonable gain, there is considerable variability. See text.

In patients with poor vision, such as those persons who are not wearing their glasses or contacts during the testing, or more rarely a person with a with retinal pigmentary degenerations, from time to time the eyes may "get lost" during tracking. Then they may show a characteristic pattern of searching saccades. However, because the patient can find the target intermittently, numerical figures for tracking may be normal. The figure above shows an example of such a case. Note the flurry of square wave jerks in the horizontal trace (top), and also the wavering around of the eye position. This patient numerically tracked with a gain of 0.96 (which is normal).

The term "disorganized pursuit" is sometimes applied to severely abnormal pursuit falling into one of the categories mentioned above. This can be a severe problem with video systems which do not allow corrected vision (i.e. .wearing of glasses during the test). Part of doing the oculomotor part of an ENG should be to measure visual acuity in the eye that is viewing.

Disconjugate eye movements may rarely occur during pursuit. In most cases it is not necessary to scrutinize the pursuit trace, because the same underlying disorders which cause disconjugate pursuit, also cause disconjugate saccades.

Pursuit in parietal lobe lesions

Individuals with parietal lobe lesions can be inattentive to half of their visual space. For this reason, pursuit may simply "stop" when the eye moves into the field where the person is not attentive (see image below).

parietal pursuit


We have moved our section on OKN/OKAN to this page.

General References

Copyright October 6, 2013 , Timothy C. Hain, M.D. All rights reserved. Last saved on October 6, 2013