Disorders of Saccadic Accuracy

Timothy C. Hain, MD• Page last modified: May 1, 2022

Of this list, overshoot dysmetria is important but rare. It is something that the operator of an ENG machine must be able to recognize - -commercial ENG scoring is nearly useless. We hope that eventually the manufacturers of ENG software will produce a useful product. The other disorders listed here are either unimportant, extremely rare, or extremely vulnerable to technical artifact.

The table above lists the most common patterns of saccadic inaccuracy which include overshoot dysmetria, undershoot dysmetria, glissades and pulsion. These disorders are caused both by ocular disorders and central nervous system disorders.

There are several pitfalls to be aware of when considering the diagnosis of dysmetria.

Blink artifact is the most troublesome because many subjects blink with every saccade, unless otherwise instructed. Blink artifact can be easily seen in figures 4 and 6 where there are brief deflections in the vertical trace, lasting about 200 msec, accompanied by synchronous deflections in the horizontal traces. Blinks contribute a technical artifact due to interactions with the EOG and infrared methods of measuring eye movements. Only the magnetic scleral eye coil technique of measuring eye movements is immune to blink artifact. EOG recordings are mainly affected in the vertical lead, but in infrared recordings, both the horizontal and vertical components are affected. When using EOG recordings, it is quite common for the direction of blink artifact to differ between each eye, or for blink artifact to be strong in one eye, and absent in the other. These problems are usually related to errors in electrode placement. Blinks are also accompanied by a small eye movement (Riggs et al, 1987), and also may interact centrally with saccades causing overshoot (Hain et al, 1986). Blink artifact is best avoided by having a vertical lead recording available, which allows one to ignore saccades with superimposed blinks, and by instructing the patient to avoid blinking during the testing. When a vertical lead is not available, such as in figures 2 and 7, it is quite difficult to be sure that a saccade of unusual configuration is truly aberrant, and one may have to fall back on direct visual inspection of the patient.

A more subtle pitfall relates to calibration error. Certain commercial electronystagmography systems calculate metrics by comparing the actual saccade displacement to the target displacement. In this situation, an incorrect calibration can cause a numerical dysmetria which is an artifact of the calibration error. This mistake can easily be detected by inspecting the eye position traces, as true dysmetria is always accompanied by corrective saccades.

A third pitfall is loss of spectacle correction. People who can't see are not accurate in following little dots on the screen. When you remove the patient's glasses to do the test, you may also remove their vision.

Figure 4: Horizontal overshoot dysmetria. Blue is target, green eye, lower red, vertical. The eye often overshoots the target. This patient had a cavernoma of the left middle cerebellar peduncle bleed and then undergo resection. There is no overshoot for vertical saccades.


Overshoot dysmetria

In overshoot dysmetria, the initial horizontal saccade is too large and the corrective saccade occurs in the opposite direction to the target displacement. Figure 4 shows overshoot dysmetria in a patient with a cerebellar lesion.

Overshoot dysmetria is not always abnormal. In normal subjects, transient overshoot dysmetria is common in saccades directed towards primary position, in saccades less than about 10 degrees in size, and saccades made to a stimulus appearing in a novel location. Normal subjects, however, will readjust their saccades to a predictable target location and, after several refixations to the same place, stop producing overshoots. Overshoot dysmetria is abnormal when it is frequent (at least 50% of the time), of significant size (greater than 2 degrees), and when it occurs in centrifugal saccades larger than 20 degrees. While numerical criteria for overshoot are available Weber and Daroff (1971), we do not feel these are necessary, as the diagnosis is usually obvious from inspection.

Axial and Saggital views of cerebellum with metastatic tumor from breast involving the vermis. The tumor is the white irregular area in the center of each picture. This patient had profound saccadic dysmetria.


saccadic vermis
Saccadic dysmetria in patient with vermal lesion (image courtesy of Dr. Dario Yacovino).
Supplemental material : Video of saccadic dysmetria (same patient as above).

Enduring overshoot dysmetria is a classic sign of a cerebellar lesion (Selhorst et al, 1976; Ritchie, 1976).

It can also occur in the abducting eye in internuclear ophthalmoplegia, in patients with visual field disturbances, and in the stronger eye of a habitual paretic-eye fixator.

  • Komiyama A, Toda AH, Johkura K. Edrophonium-induced macrosaccadic oscillations in myasthenia gravis. Ann Neurol 1999:45:522-525

Undershoot dysmetria-- a common oculomotor "abnormality"

In undershoot dysmetria, the initial saccade is too small and the corrective saccade continues onward towards the target. Undershoot dysmetria, at least to both the right and left, does not carry the same pathologic connotation as does overshoot dysmetria as undershoot is common in normal subjects. Normal subjects will show about 1-2 degree of undershoot for 20 deg and larger target displacements (Lemij and Collewijn, 1989). Undershooting of saccades is very common in clinical contexts -- patients with poor vision, such as due to cataract or inability to wear glasses during oculomotor testing, may simply be guessing as to new target location, and can produce undershoot or overshoot patterns, typically to both sides. For this reason, undershooting is not specific sign of any particular disease and not terribly much should be made of it on ENG testing.

Constant and significant (first saccade < 50% of target displacement) undershooting is suggestive of a basal ganglia disorder such as Parkinson's disease or progressive supranuclear palsy (PSP), and degenerative nervous system diseases such as Tay-Sachs (Rucker et al, 2004).

Undershoot to one side only is suggestive of a visual field defect or a unilateral cerebellar lesion. The figures above and belowshows an example of hypometric saccades produced by patients with left hemianopia. Note in the top that there is a mixture of undershoot and overshoot, and that the inaccuracy is mainly to the left. Saccades are of normal velocity. In the bottom, note that there are no overshoots, but there are searching saccades to the left.

homo hemi


Staircase saccades

On rare occasions, people will make staircases rather than saccades or slow-phases.


We mainly see these in patients who are sedated. We are dubious that they are the consequence of a "disease". The patient above had no neurological disorder.

Staircase saccades

Well anyway, staircase saccades have been reported in a variety of central disorders -- including Parkinsonism and cerebellar ataxia. In theory, this should be related to a disorder of central saccadic processing in the brainstem. A command is given to the superior collicullus, and it stops too soon, requiring a restart and a staircase. The patient illustrated above had an unknown disorder with severe ataxia but normal inner ear function.

Near Blindness:

Surprisingly, we often forget that if blind people are totally inaccurate, nearly blind people may also have problems following little dots around on screens. Furthermore, when we remove people's glasses, we may be making them nearly blind.

An interesting example of this is shown in the saccadic trace below, from a patient with Retinitis Pigmentosa. Persons with RP eventually develop tunnel vision. Thus, when the dot is moved for the saccade paradigm, a person with RP has to find it, as in essence, it has simply disappeared.

RP saccades
Dysmetric saccades in persons with retinitis pigmentosa (RP), which is an ocular disorder which impairs peripheral vision. Latency is long, saccades both under and overshoot or overshoot, depending on the style of the patient who is tracking.


Because central vision is spared in RP, smooth pursuit may be normal. This is illustrated on the page that discusses tracking disorders.

Pulsion during vertical saccade test Horizontal no pulsion
Vertical saccades (blue) are accompanied by horizontal displacements (red trace) A technical error is not the cause as there are no vertical movements for horizontal saccades.

Patients who are hemianopic due to an occipital territory stroke cannot see targets in their blind hemifield but while their verbal detection reports were at chance level, they still can make inaccurate and longer latency saccades into their blind fields. This is called "action-blindsight". (Fayel et al, 2014). This could be contrary to what one would expect from a bedside exam.


The term "pulsion" is applied to vertical saccades that are pulled to the right or left, requiring a horizontal corrective saccade to fixate the target. Both upwards and downwards saccades are pulled in the same horizontal direction. Pulsion towards the side of lesion, or "ipsipulsion", occurs after infarcts in the distribution of the posterior inferior cerebellar artery ( Kommerell and Hoyt, 1973; Meye et al, 1980). Pulsion away from the side of lesion, or "contrapulsion", may occur after infarcts in the distribution of the superior cerebellar artery (Ranalli and Sharpe, 1986). Most clinical laboratories do not attempt to record pulsion, although current equipment is capable of resolving pulsion.

Lateropulsion was used earlier to describe a different situation where the eyes deviate to one side under closed eyelids (Hagstrom et al, 1969). This is a confusing situation because of the possible confusion between sacacdes and eye deviation under closed eyelids. We suggest using the term "eye deviation under closed eyelids", rather than "pulsion".

When diagnosing pulsion, the main technical error is due to the camera being turned so that there is cross-coupling of horizontal into vertical and vici-versi. This can be disproven by noting that the inappropriate "pulsion" of the eye is followed by a corrective saccade (or a drift), and that vertical saccades have no horizontal movement.

  • Hagstrom, L., et al. (1969). "Oculostatic and visual phenomena occurring in association with Wallenberg's syndrome." Acta Neurol Scand 45(5): 568-582.
  • Kommerell, G. and W. F. Hoyt (1973). "Lateropulsion of saccadic eye movements. Electro-oculographic studies in a patient with Wallenberg's syndrome." Arch Neurol 28(5): 313-318.
  • Meyer KT, Baloh RW, Krohel GB, Dow BM. (1980) Ocular lateropulsion: a sign of lateral medullary disease. Arch Ophthal 98, 1614-1616
  • Ranalli PJ, Sharpe JA. (1986) Contrapulsion of saccades and ipsilateral ataxia: A unilateral disorder of the rostral cerebellum. Ann Neurol 20: 311-316


The term "glissade" designates a saccade which does not end crisply, but rather glides to its end point. "Onward glissades" occur when the eye continues to glide in the same direction as the faster part of the saccade, and "backward" glissades occur when the eye drifts in the opposite direction as the main saccadic movement. Figure 6 illustrates backwards glissades in a patient with myasthenia gravis. Glissades occur in conditions in which the brainstem miscalculates the "pulse" of oculomotor activity needed to get the eye to new position or the "step" of innervation needed to hold the eye in place against elastic forces. Thus glissades are often said to be due to a "pulse-step mismatch". Patients having rapid changes in oculomotor function, such as ocular myasthenics are particularly prone to developing a glissadic pattern, because the amount of neural firing required to obtain a given eye position and to hold it there against elastic restoring forces is constantly varying. Myasthenics also may demonstrate a briefer drift called "quiver" (Yee et al, 1976). Quiver does not occur in Eaton-Lambert syndrome (Dell'Osso et al, 1983). Patients with cerebellar lesions may produce glissades because they are unable to adjust their pulse step ratio. Patients with internuclear ophthalmoplegia show onward prolonged glissades in the adducting eye, and briefer backward glissades in the abducting eye.

The main pitfall to consider when trying to decide if a patient has glissades is the adequacy of head stabilization. If the head is free to move and does so during a saccade, the eye-component of a combined head-eye saccade may resemble a glissade. Infra-red recordings also have a special problem as they may show a glissade-like artifact related to changes in eyelid position which accompany saccades.