Journal of Neurology, Neurosurgery, and Psychiatry 1982;45:431-439

Acquired pendular nystagmus: its characteristics,localising value and pathophysiology

MA GRESTY, JJ ELL, LJ FINDLEY

From the Medical Research Council, Neuro-Otology Unit, Institute ofNeurology, The National Hospital,London, UK

SUMMARY Investigations were made of 16 patients with acquired pendular nystagmus and afurther 32 cases reported in the literature were reviewed. Amongst our own patients two thirdshad multiple sclerosis, almost one third a cerebrovascular accident or angioma and two had opticatrophy with squint. The nystagmus took forms which could be monocular or binocular, conju-gate or disconjugate and could involve movements about single or multiple axes. Spectral analysiswas used to characterise the amplitude and frequency of the movements and to estimate thedegree of relationship (coherence) between movements of the two eyes or between movementsof one eye about several axes. The oscillations ranged in frequency from 2-5 Hz to 6 Hz, withtypical amplitudes between 30 and 5°. In a given patient all oscillations, regardless of plane, werehighly synchronised. Somatic tremors of the upper limb, face and palate associated with thenystagmus were often at similar frequencies to the eye movement. The other ocular signs com-mon to all our patients were the presence of squint with failure of convergence. Most patients alsohad skew deviation or internuclear ophthalmoplegia or both. The major oculomotor systems, thatis, saccades, pursuit, optokinetic and vestibulo-ocular reflexes could be intact. It is inferred thatthe mechanism responsible for the pendular nystagmus lies at a level which is close to theoculomotor nuclei so that it can have monocular effects but is not part of the primary motorpathways. It is possible that this mechanism normally subserves maintenance of conjugatemovement and posture of the eyes. The periodicity of the nystagmus is likely to arise frominstability in a certain type(s) of neurone, for the associated somatic tremors have similar charac-teristics and yet involve very different neuronal muscular circuitry. Prognosis for cessation of thenystagmus is poor. In five patients with multiple sclerosis it was suppressed by intravenoushyoscine with, however, unacceptable subsequent side effects.

"These eyes that rowle in vainTo find thy piercing ray, and find no dawn."

John Milton, Paradise Lost, III, 11/23-24

Pendular nystagmus is a sinusoidal, involuntaryoscillation of the eyes which may be congenital oracquired in nervous diseases. Clinically, differentia-tion between congenital and acquired nystagmus isnot difficult because of their distinct characteristics.Congenital pendular nystagmus is horizontal,binocular and conjugate, can have a variable wave

Address for reprint requests: Dr MA Gresty, the Institute ofNeurology, Queen Sq, London WC1N 3BG, UK

Received 6 November 1981. Accepted 17 January 1982

form and changes into a jerk nystagmus on lateralgaze. Acquired pendular nystagmus may be mono-cular or binocular, conjugate or disconjugate with awave form which is largely independent of eye posi-tion. The movement of the eye may be vertical,horizontal or torsional in which case the globerotates about the principal axis. An observedmovement may be the result of combined oscilla-tions about more than one axis as in "rotary" nys-tagmus (fig 1). Pendular nystagmus sometimesoccurs in association with rhythmical movements atsimilar frequencies of other parts of the body, as inthe syndrome of "oculo-palatal-myoclonus"l inwhich postmortem examination has confirmed alesion in the brain stem. The nystagmus in this syn-

431

Protected by copyright.

on 6 August 2018 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.5.431 on 1 May 1982. D

ownloaded from

Gresty, Ell, Findley

Trajectory of frontal pole(within 50% of actual size)

Saccade -

Horizontalpendularcomponert

Blink

Vetcalperwularcomponent

Fig 1 An example ofmonocular rotatory nystagmusrecorded from a 25-year-old male patient with multiplesclerosis. The rotation was the result ofsynchronouspendular oscillations in both the vertical and horizontalplanes. The central figure indicating the trajectory wasproduced by applying the horizontal and verticalelectro-oculographic recordings to the y and x amplifiers ofan oscilloscope. The maintenance ofthe unchangingelliptical trajectory indicates the precision with whichhorizontal and vertical components were synchronised.

drome is always similar to the form which can occurin isolation and there is no reason to suspect thatthey have separate origins.

Although the planes in which the eyes move andthe phase relationships between the movementsshow considerable variability from one case toanother there are features common to acquiredpendular nystagmus which lead us to suspect thatthe movements should be classified together. Firstly,both within and between patients the frequency andamplitudes of the nystagmus are similar giving a clin-ical appearance which is quite distinctive regardlessof the plane of movement. In addition, in any indi-vidual, nystagmus which occurs in disconjugateplanes is synchronised indicating some form of yok-ing between vertical, horizontal and torsionaloculomotor systems. These features suggest that aunitary mechanism or at least similar types ofmechanisms sharing a common rhythm generatorare responsible for the various manifestations ofpendular nystagmus.

In the only extensive survey of acquired pendularnystagmus to date2 the authors postulated that theocular movements originated from the cerebellarnuclei and their connections and were evidence of afailure to maintain eye posture. The existence ofmonocular pendular nystagmus throws doubt uponthis theory.3 Accordingly we examined 16 patientswith acquired pendular nystagmus with a particularview to determining the abnormalities of ocularmovements common to all which might provideclues as to aetiology and pathophysiology.

Methods

Recordings of eye movements were taken using directcoupled electro-oculography with a low pass cut off usuallyset at 32 Hz. In cases of torsional pendular nystagmus twosmall silver-silver chloride electrodes were placed on theupper lid over the insertion of the superior obtique muscleinto the globe. The signals were amplified with a widebandpass in order to detect the EMG activity of the obli-que muscle. Accelerometric recordings of head or limbmovements were taken using linear piezo resistiveaccelerometers (Endevco 7265-10) mounted with theirsensitive axes oriented in the major plane of movement.The device had a frequency response from steady stateacceleration up to a superimposed low pass cut off at32 Hz. Recordings of limb displacements were obtainedusing a Schottky barrier photo-detector mounted with itssensitive surface in the focal plane of a 35 mm camerawhich was oriented to view the limb from a suitable angle.A small pilot lamp was mounted on the limb, the move-ment of which was registered by the photo-detector andresolved into displacements in the vertical and horizontalplanes (after the method of Findley et al4).The signals relating eye and body movements were sub-

jected to spectral analysis on a Hewlett Packard 5420Adigital signal analyser. The analyses employed to character-ise the signals were the Auto-spectrum; used to measurethe amplitude and frequency of the dominant frequency ofa tremor, and Coherence (the average of square of thecross spectra divided by product of the averages of theauto-power spectra); which is a measure of the degree towhich two signals (that is, tremors) are related. Coherenceis expressed on a scale of 0 to 1 in which 0 indicates thatthe signals are independent of each other and 1 indicatesthat they are completely related. As a point of interpreta-tion a coherence of 1 can be taken to indicate that thesignals are causally related. These functions derived fromthe raw signals were averaged to improve the significanceof the measurements. The analyses were performed on1025 point digitised 10-24 s samples of raw data producinga maximum temporal resolution of 10 ms, a frequencyresolution of 0-05 Hz and bandwidth of DC to 25 Hz. Thedata was digitally filtered to prevent aliasing of frequenciesin excess of 25 Hz and windowed with a Hanning windowappropriate for the analysis of sinusoidal signals.5 Aver-ages were made on spectra derived from overlapping sam-ples of raw data. This compensates for the reduced con-tribution of the data at the beginning and end of the datasamples which results from the application of a Hanningwindow.The validity of the raw data signals was established by

displaying the signals on a storage oscilloscope and makinga simultaneous comparison with an infra-red video record-ing of the actual eye movement. Combined vertical andhorizontal movements were plotted in the form of a Lissaj-ous figure using the x and y amplifiers of the oscilloscope(fig 1).Patients

Sixteen patients were referred for assessment by the con-sultants at The National Hospital, Queen Square and

432

Protected by copyright.

on 6 August 2018 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.5.431 on 1 May 1982. D

ownloaded from

Acquired pendular nystagmus: its characteristics, localising value and pathophysiology

St Thomas' Hospital, London. In addition we include dataderived from a further 27 patients with multiple sclerosisreported by Aschoff et al,2 four patients described byChokroverty and Barron6 and one of Nathanson7 and fourtaken from the case history records of the National Hospi-tal who had had electro-oculographic recordings taken oftheir eye movements; a total of 52 patients with monocularand binocular pendular nystagmus. Congenital nystagmusof the pendular variety was excluded from the survey.Amongst our patients 12 had clinically definite multiple

sclerosis; four had had brain stem vascular accidents; one abrain stem angioma; there was one patient with congenitaloptic atrophy with strabismus and one with a unilateralamblyopia secondary to strabismus (both of these patientshad a uniocular vertical pendular nystagmus in the affectedeye); and one patient had a provisional diagnosis of brainstem encephalitis.A universal complaint amongst these patients was one of

oscillopsia which was usually sufficiently marked to impairreading ability. The majority of our patients had severeneurological disability but in the few who were ambulantand active the oscillopsia caused difficulty with dailyactivities. The oculomotor impairment common to allpatients was some form of failure of conjugate gaze ordefect of voluntary convergence.

Results

The relevant findings in our patients which relate tothe behavioural characteristics of their pendular nys-tagmus, associated oculomotor and visual disorders,tremor of parts of the body and diagnosis are pre-sented in the table.

FREQUENCY OF PENDULAR NYSTAGMUSThe histogram of frequencies of pendular nystagmusas measured in our own patients and as reported byother authors is presented in fig 2. In the cases ofnystagmus which varied in frequency the middle ofthe range of variability was used to plot the histo-gram. Because of the finding that in any one patient,regardless of the plane of movement, pendular nys-tagmus in more than one direction was phaselocked, such compound nystagmus was character-ised by only one frequency measurement. The fre-quency distribution was unimodal, skewed towardslower frequency values, with a mode at between 2-6and 3 Hz with a mean at 3-6 Hz. There was no obvi-ous relationship between frequency and direction ofnystagmus and an underlying disease process.

In any given patient, whenever an episode of welldefined binocular pendular nystagmus occurred,there was a precise phase lock between the move-ments of the right and left eye. The phase relation-ship was independent of the plane of movement,which could be different in the two eyes. Thischaracteristic is illustrated in the case of a 28-year-old female with a presumed cerebrovascular acci-

dent which had left her with a binocular convergentpendular nystagmus which had remained unchangedin character for two years. An analysis of her eyemovements is presented in fig 3. The raw datarecords in the lower part of the figure show afluctuating pendular nystagmus of slightly greateramplitude in the left eye. The synchronisation arrowbetween the traces indicates that the nystagmus con-sisted of pendular movements of convergence anddivergence. The auto-spectra of the movements ofthe left and right eye are averages of ten overlapping10-24 s samples of eye movement recordings andindicates a frequency of 3.4 Hz in both eyes. Thecorresponding coherence function measured on thesame data sample as the auto-spectra indicates acoherence of unity in the frequency band of thependular eye movement (indicated by the single ver-tical arrow). The phase relationship between theright and left eye measured over the same data sam-ple was 150° indicating that the movements were notperfectly convergent but slightly staggered (a phaseof 1800 would indicate perfect convergence). Themost important of these measurements is the coher-ence function which increases its specificity andsignificance with successive averaging, unlike theother measurements in which averaging can hide thevariability of the phenomena. The coherence ofunity in the frequency band of the pendular move-ments proves that they are frequency locked.

Similar measurements on other patients withcompound nystagmus revealed a similar degree ofrelationship between movements in the variousplanes. In the cases of combined horizontal and ver-tical oscillations the eye executes a rotation (as infig 1) in which the circular or eilliptical pattern ofthe rotation is invariable indicating that frequencyand phase relationships between the horizontaland vertical components of the movement are alsoinvariable.

CONCURRENT TREMOR OF OTHER PARTS OF THEBODYEight patients with pendular nystagmus on whomwe made measurements had tremor of other parts ofthe body including the head, neck and upper limbs.Of such tremors, some were similar in frequency tothe pendular nystagmus and others dissimilar. Bothsimilar and dissimilar tremor frequencies couldcoexist in the same person and in the same limb(depending upon whether the tremor was broughtout by a posture of flexion or extension). The histo-gram of somatic tremor frequencies in our patients ispresented in fig 3. It is evident that the distributionof frequencies is similar to that of the pendular nys-tagmus.The degree of synchronisation which can coexist

433

Protected by copyright.

on 6 August 2018 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.5.431 on 1 May 1982. D

ownloaded from

Table Classification ofdiseases, ocular signs and associated somatic tremor in patients with pendular nystagmus

Age Sex Pendular nystagmus Frequency phase Modifying factors Associated tremors INO Convergence Other oculomotor signs

direction andamplitude

Right eye Left eye

Multiple sclerosis37 F H 40 H 40 3-6 Hz 330 absent Opsoclonus, 1st° nystagmus

to right and left

43 F H 20 H 20 3-5 Hz Fluctuating amplitude Intention/postural absent Exophoria, centeringprovoked by saccades arm tremor 3-5 Hz dysmetria upbeat nystagmus

24 M H 50 H 2° 35Hz 500 bilateral absent

60 M H 40 H 40 4-3 Hz Accentuated on near bilateral poorgaze

23 M V V 2-8 to Wide fluctuation in Head, palate, labial bilateral absent3-6 Hz amplitude/frequency 2-8-3-6 Hz

28 M H 30 H 30 3-7 to Wide fluctuations Head 3-7 Hz absent Supranuclear palsy, up/

acuities 2/60 4-0 Hz down beat nystagmus,bilateral 6ths

25 M R 40 T 40 3-2 Hz 850 Absent on closure bilateral poor ? Incomplete INO, 1stonystagmus to right and left

24 M R 30 H 30 3-7 Hz 1020 Episodic, provoked Head 3-7 Hz absent Supranuclear palsy, up/by saccades down beat nystaFmus

43 F H H 3.5 Hz Postural arm absent 1sto nystagmus nght andtremor 3-5 Hz left, upbeat nystagmus

49 M H 20 3-8 Hz bilateral ? Records: incomplete data

45 F F + H 40 3-7 Hz bilateral ? Records: incomplete data

54 F F 20 6-0 Hz Absent in darkness ? ? Records: incomplete dataand on closure

Cerebrovascular accident49 F V H 2-5 Hz 600 Rubral tremor of bilateral absent Skew deviation, omni-

right arm 2-5 Hz directional gaze pareticnystagmus

56 M V 80 V 80 2-4 Hz Persisted through absent Total horizontal gazelight anaesthesia paresis, skew deviation,

downbeat nystagmus

51 M V + T 2-50 minimal? 2-8 Hz Palate 2-5 Hz unilateral absent

28 F H 40 H 40 3-5 Hz 1500 Variable incidence poor Square wave jerks,monocular jerk nystagmusof right eye

Brain stem angioma51 F V 40 V 40 2-2 Hz bilateral absent Skew deviation

Brain stem encephalitis64 F V 80 V 80 2-7 Hz Enhanced with eye Thumb, palate bilateral absent

closure, variable 2-7 to 3-5 Hz

Amblyopialoptic atrophy40 M V3b 2-9 Hz Intermittent, absent poor Right exotropia and

amblyopic right eye in dark, provoked skew deviationby fixation

9 M V 70 V 7° 3-5 Hz Left eye nystagmus Left eye optic atrophyonset before right esotropia of left eye until 8

year subsequently becameexo

H=horizontal, V=vertical, T=torsional.

between pendular nystagmus and simultaneous tre-mor of other parts of the body is illustrated in fig 4which presents raw data records from a 23-year-oldmale patient with multiple sclerosis. His nystagmusranged in frequency from 2-8 to 3-6 Hz and was

accompanied at various times by rhythmical move-

ments of the head, upper lip or palate. When any

combination of these occurred simultaneously theywere approximately synchronous as indicated by thevertical synchronisation arrows in fig 4. Carefulmeasurement of the traces, however, revealed tlatover several cycles the various tremors did not main-tain a precise phase relationship.

EFFECT OF PENDULAR NYSTAGMUS ON VISUALACUITYAll patients were distressed by the oscillopsia whichthey experienced as a result of their pendular nys-tagmus and this was more or less proportional to theamplitude of the eye movement. We were, however,surprised that visual acuity in patients with pendularnystagmus did not seem to be particularly low whentested using a Snellen chart, as estimated by com-parison of acuities measured before and after treat-ment with hyoscine to suppress the nystagmus,acuity was reduced by only one to two lines of theSneilen chart. The degree of visual impairment

Gresty, Ell, Findley434

Protected by copyright.

on 6 August 2018 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.5.431 on 1 May 1982. D

ownloaded from

Acquired pendular nystagmus: its characteristics, localising value and pathophysiology 435

Frequency of ocular oscillaton which our patients experienced seemed more the13 result of oscillopsia than of loss of acuity.112 * Aschoff Severe problems were encountered when the10 * Chokroverty patients attempted to read when it became evident9 * *Records that the nystagmus interfered with the generation of8 * Preset normometric scanning steps. The mechanism6 _ appeared to be one of simple summation. When thes+ ., + patient attempted to generate a saccade the amp-3 * litude was modified by the addition of the involun-2 | tary oscillation which could either reduce or aug-

2 3 4 5 6Hz ment the saccade depending upon the phase of theI

nystagmus.

3 BEHAVIOURAL CHARACTERISTICS4 As can be seen from the column in the table listing6 Associated somatic tremor modifying factors in the pendular nystagmus mani-

fest in our patients there were no consistentFig 2 Frequency histograms ofthe frequencies ofpendular behavioural characteristics which could be helpful innystagmus and associated somatic tremor in our own determining the underlying aetiology. In particularpatients and in patients reported by other authors and those we did not observe any nystagmus to be inhibiteddocumented in the records ofthe National Hospital. following saccadic eye movements as reported by

Right eye Auto spectra Coherence30 . . 0*9

GO Hz 10'0 00 Hz 1001800-

~~Phase30

0<0 Hz 10<0 0-0 Hz 100

Right eye

5 [ Left eye monocular saccades

ls

Fig 3 Binocular pendular nystagmus in the horizontal plane recorded from a 28-year-old female patient with apresumed cerebrovascular accident. The movements ofher eyes were in the direction ofconvergence and divergence.The auto spectra indicate the common dominant frequency ofthe pendular movements. The coherence and phasemeasurements at this frequency (indicated by small vertical arrows) show that the movements are completelysynchronous with a phase difference of 150°. The raw data records show two types ofspontaneous involuntaryactivity. The pendular nystagmus which is common to both eyes as indicated by the synchronisation arrow andmonocular saccadic movements ofthe right eye. The spectral measurements were based on averages derived from ten1024 s samples ofdata. On this number ofaverages a coherence of0.7 was significant at the 2% level.

Protected by copyright.

on 6 August 2018 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.5.431 on 1 May 1982. D

ownloaded from

Gresty, Ell, Findley

Up

Down ,

Pflate Palate

Head

ls Upper lip

Fig 4 Recordings ofsomatic tremor and pendular nystagmus taken froma 24-year-old male patient with multiple sclerosis. Synchronisation arrowsbetween the traces indicate that the movements are similar in frequency inall body parts.

Aschoff et al.2 On the contrary we found a curiousinstance of an episodic pendular nystagmus in a24-year-old male patient with multiple sclerosiswhich was provoked by saccadic movements of smallamplitude as illustrated in fig 5. With attempts tohold the eyes still there were short bursts of oscillat-ory eye movement. Whenever the patient madesmall saccades of about 5° to 10° in amplitude two orthree cycles of nystagmus were initiated whichdecayed in the form of a damped oscillation. Therewas an accompanying tremor of the head with com-parable behavioural characteristics. Thecrescendo-decrescendo appearance and the initia-tion by sudden movement of such rhythmicalinvoluntary movements in this patient are reminis-cent of "startle myoclonus" reported in children.8

It has been said that acquired pendular nystagmusin multiple sclerosis is present only during fixationand ceases with eye closure. We found this to be thecase in only two of our multiple sclerosis patients.

PHARMACOLOGICAL SUPPRESSION OFPENDULAR NYSTAGMUSOn the basis of the observation that an intravenousdose of hyoscine can temporarily suppress low fre-quency somatic tremor in certain cases of multi-system degeneration,9 we investigated the effect ofthis drug in five patients with pendular nystagmus.Four of these had multiple sclerosis and one a brainstem cerebrovascular accident. In all five,10 minutes after a single intravenous dose of hyos-cine 0-4 mg, the nystagmus was abolished (as judgedby observation and electro-oculography). Thepatients themselves were aware of this and no longerexperienced oscillopsia. In one case visual acuityimproved from 6/12 to 6/6. The nystagmus recurredafter approximately 15 minutes. The patients subse-quently noted slight drowsiness but did not experi-ence this during the period when the nystagmus was

absent. The potential long term therapeutic applica-tion of hyoscine in such cases is being further inves-tigated.

R Spontaneous oscillation

L

Induced oscillationR .

L

isFig 5 Bursts ofpendular nystagmus recorded from a24-year-old male patient with multiple sclerosis. When thepatent tried to keep his eyes still there were bursts ofcrescendoldecrescendo pendular oscillations as in the uppertraces. Execution ofsmall saccades appeared to trigger offtwo or three cycles ofnystagmus as shown in the lowertraces.

In several other patients, a range of drugs includ-ing leva-dopa, baclofen, clonazepam, prochlor-perazine, carbamazepine, tetrabenazine and a vari-ety of major tranquillizers were administered to noavail. Informed consent following a full explanationof the procedure was obtained from all patients withwhom pharmacological suppression of nystagmuswas attempted.

ASYMMETRICAL FEATURES OF THE WAVE FORMOF PENDULAR NYSTAGMUSThe electro-oculographic technique which was theonly eye movement recording method available tous is not sensitive enough to transduce the waveform of pendular nystagmus of small amplitude witha signal to noise ratio adequate to reveal the shapeof the movement trajectory in detail. Informationabout precise trajectory would be useful because itwould indicate whether the muscular activityresponsible was reciprocally organised or asymmet-rical in time or amplitade, which could give clues asto the origin of the oscillation. However, two obser-vations derived from electro-oculographic record-ings did suggest that in some cases the trajectory ofthe movement was asymmetrical indicating that

436

Protected by copyright.

on 6 August 2018 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.5.431 on 1 May 1982. D

ownloaded from

Acquired pendular nystagmus: its characteristcs, localising value and pathophysiology

either the rhythm of the agonist/antagonist activitywas iambic or that active myogenic displacementwas occurring in only one direction with movementin the opposite direction being under the influenceof the passive elastic properties of orbital tissue.The first of these observations was that the shape

of the wave form of the nystagmus tended to be"saw tooth" rather than sinusoidal; the second wasthat in some cases the pendular nystagmus exerted adirectional bias on other oculomotor responses. Thislatter point is illustrated in fig 6(a). The recordingsare taken from a 25-year-old male patient with afour year history of multiple sclerosis. He had a rota-tory nystagmus of the right eye with a torsional pen-dular movement of the left. The movements of hisright eye in response to optokinetic simulation(OKN) using a rotating drum which provided fullfield stimulation were markedly deranged with drumrotation to the right and relatively preserved withrotation to the left. His responses to caloric androtational stopping stimuli were similarly affected.The movements of his left eye were normal. Smoothpursuit movements of the right eye were similarlyasymmetrical with the oscillations tending to add tothe pursuit movement in the rightwards directiongiving a hypermetric appearance and to subtractfrom the pursuit in the leftwards direction giving ahypometric appearance. Corresponding movements

A OKN

Righteye

R

Left eye

0 \20B/s L2B /

Pursuit

Nov 1980 ls

R5°] Feb 1981L-,

lsFig 6 (a) recordings taken from a 25-year-old male patientwith multiple sclerosis with a monocular pendularnystagmus ofthe right eye. The nystagmus interferes withoptokinetic nystagmus to the right producing a derangedresponse. Nystagmus to the left and nystagmus in the goodeye is unaffected. Smooth pursuit in the right eye was moreaffected with rightwards than with leftwards movements(arrow). (b) successive records taken at a 4 month intervalfrom a 49-year-old female patient with a cerebrovascularaccident showing an alteration in the nystagmus wave formfrom pendular to triangular.

of the left eye showed a mild pursuit deficit. Theshape of the pendular movements of the right eyewhen the patient was trying to keep his eyes stillsuggested that there was a stronger pull to the rightthan to the left. If there were such an asymmetry ofmuscle action then it would explain the directionalcharacteristics of the patient's pursuit, vestibularand optokinetic responses.

Further evidence for asymmetry in the mechanismof some pendular nystagmus is derived from anotherof our patients; a 49-year-old woman with a brainstem cerebrovascular accident who presented origi-nally with a pendular nystagmus and other signs of abrain stem lesion. Over the course of four monthsthe signs in her upper limbs partially remitted. Con-currently the pendular nystagmus became a "jerk"nystagmus as illustrated in fig 6 (b).The implications of these observations upon the

interpretation of the results of caloric and rotationaltests are important. The possibility of an asymmetri-cal wave form in pendular nystagmus should betaken into account in interpretation of such tests.Failure to do so could easily lead to a misleadingclinical interpretation.

Discussion

PHYSIOLOGICAL AND ANATOMICALCONSIDERATIONS CONCERNING THE NATURE OFPENDULAR NYSTAGMUSAschoff et a12 postulated that a lesion in the cerebel-lar nuclei or their connections was the cause ofacquired pendular nystagmus. Their reasons for thisincluded the fact that amongst their patients withpendular nystagmus (all of whom had multiplesclerosis), many had signs of cerebellar nuclearlesions in their upper limbs, for example postural orintention tremor. In addition they reviewed the evi-dence concerning the site of lesions in oculo-palatalmyoclonus which consistently include the dentateand inferior olivary nuclei and brachium conjunc-tivum.7 10 They interpreted the pendular eye move-ments as reflecting loss of postural control of theeyes by the cerebellar nuclei.'2 There are severalreasons why this view is difficult to maintain. Firstly,as emphasised by Castaigne et al,3 the existence ofmonocular pendular nystagmus (in the case of Cas-taigne's patient a monocular circumduction nystag-mus, that is, rotatory nystagmus) implies that themechanism responsible for generating the oscillationmust have access to an oculomotor nucleus or nucleiwhich innervate only one eye. Although theinterpretation of stimulation experiments is difficultbecause of the possibility of the excitation of axoncollaterals, the evidence derived from animal andhuman studies indicates that electrical stimulation of

437

Protected by copyright.

on 6 August 2018 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.5.431 on 1 May 1982. D

ownloaded from

438

the cerebellum, its nuclei or efferent pathwaysresults in binocular movements." '1314 Secondly, it isimprudent to argue that the coexistence of pendularnystagmfis and somatic tremor in diseases with mul-tifocal lesions implies that these signs arise from thesame anatomical lesion. This point is highlighted bythe fact that amongst the patients in Aschoff's sur-vey there was by no means a one to one correspon-dence between cerebellar signs in the upper limbsand the occurrence of pendular nystagmus and thatof our own patients less than half had unequivocalsigns of cerebeliar disease.The unique feature of pendular nystagmus which

requires explanation is that it can be in any plane orcombination of planes in either eye, simultaneouslyand at the same frequency. An important clue as tothe nature of its mechanism is that all other saccadicand slow phase oculomotor mechanisms exceptthose responsible for conjugate gaze and vergencemay be intact.Because the generation of saccadic, pursuit,

optokinetic and vestibulo-ocular reflex eye move-ments may be intact in the presence of pendularnystagmus the causal mechanism is unlikely to berelated to these major oculomotor systems. The factthat a compound pendular nystagmus may exist inwhich the movements in various planes are perfectlysynchronised (as indicated by a value of unity in thecoherence function), implies that the nystagmusarises from disorder in a mechanism which is nor-mally responsible for the co-ordination of eyemovement in more than one plane.

In any one eye movement all extra-ocular musclesare involved.'5 In order to ensure accurate and con-jugate gaze, corrections have to be made for devia-tions from the desired eye position or trajectorywhich arise from the secondary actions of musclesand changes in the passive visco-elastic properties oforbital tissue. Although not anatomically identified,it is accepted that a neuronal mechanism responsiblefor such corrections exists.'5 An important aspect ofsuch a mechanism(s) is that it frequently has a dif-ferent yet co-ordinated action on the two eyes, par-ticularly as far as the secondary action of the musclesare concerned. This type of co-ordinating mechan-ism is a candidate for involvement in pendular nys-tagmus for the consequences of damage to such astructure could explain the variety and synchronicityof eye movements encountered in the disorder.An anatomical localisation of the lesion respons-

ible for pendular nystagmus in the proximity of theoculomotor nuclei is suggested by several features ofthe nystagmus and by frequently associated abnor-malities of other ocular movements. Stimulation ofmost supra-nuclear oculomotor centres results inbinocular movement and it is only close to or at the

Gresty, Ell, Findley

level of the oculomotor nuclei that a monocularresponse is evoked.'5 It is possible that the abnormalsignals responsible for the oscillation have access tomotor neurones innervating reciprocally relatedmuscles because they produce a sinusoidal move-ment in the absence of globe retraction. In order forthis to occur the signals must arise at a supranuclearlevel. However, in the case of a monocular move-ment this must be near the third nerve nucleus.These arguments must almost necessarily hold for

a monocular nystagmus; however, they have noapplication in the localisation of a lesion responsiblefor a conjugate binocular pendular nystagmus.Further evidence that the lesion responsible forpendular nystagmus is near the oculomotor complexis to be inferred from the universal occurrenceamongst our patients of associated internuclear oph-thalmoplegia or skew deviation, failure of con-vergence and squint, all of which are signs ofderangement between or close to the level of theoculomotor nuclei.The occurrence of pendular nystagmus in mono-

cular amblyopia with strabismus requires furtherexplanation because there is no evidence of struc-tural damage in the brain stem. However, thecharacter of the nystagmus in these patients is indis-tinguishable from that seen in multiple sclerosis, forexample. The feature common to both is a markedfailure in the vergence system and this may indicatethat this is an aetiological factor. Certainly the simi-larity between nystagmus of an amblyopic eye andthat resulting from structural lesions of the nervoussystem strongly suggests that they result fromabnormalities of function of the same nervousmechanism. When pendular nystagmus arises inamblyopia we suspect there is additional, com-pounding insult to the nervous system, for example,migraine.One can only speculate on the origin of the fre-

quency characteristics of pendular nystagmus. Cluesas to the origin of the rhythm come from the factthat associated somatic tremors sometimes have asimilar periodicity. It is unlikely that such periodicityis caused by time delays around a form of controlloop because the control parameters required forbody parts with such diverse functions and physicalproperties as the eye, tongue or arm must be verydifferent. The uniformity of frequency encounteredin the syndrome of oculo-palatal-laryngeal-diaphragmatic myoclonus suggests that the rhythmi-city derives from properties which are held in com-mon by different types of neurones in differentneuronal circuits. As such these properties must befundamental to the function of the neurone, eitherin isolation or in the context of its local circuitry.Cycles of excitation and inhibition arising from the

Protected by copyright.

on 6 August 2018 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.5.431 on 1 May 1982. D

ownloaded from

Acquired pendular nystagmus: its characteristics, localising value and pathophysiology

properties of the neuronal membrane have beendescribed for example in the inferior olive by Llinasand Yarom"6 and this structure is implicated in thesyndrome of oculo-palatal myoclonus.710 None ofthe tremulous movements in the syndrome arenecessarily associated with weakness or ataxia of theaffected part which implies that the primary motorpathway is intact so that the tremor must arise froman accessory pathway.

CLINICAL ASPECTS OF PENDULAR NYSTAGMUSIn general the prognosis for cessation of involuntaryocular movement and diminution of oscillopsia ispoor. The results of pharmacological intervention sofar have been discouraging. In those cases where thenystagmus has been modified by drug treatment (forexample, with barbiturate and hyoscine) the subse-quent level of impairment of concentration would beunacceptable for everyday life.

References

Guillain G. The Syndrome of Synchronous andRhythmic Palato-Pharyngo-Laryngo-Oculo-Dia-phragmatic Myoclonus. Proc R Soc Med1938;31:1031-8.

2 Aschoff JC, Conrad B, Kornhuber HH. Acquired Pen-dular Nystagmus with oscillopsia in Multiple Sclerosis:a sign of cerebellar nuclei disease. J Neurol NeurosurgPsychiatry 1974;37:570-7.

3 Castaigne P, Chain F, Pierrot-Deseiligny C, LarmandeP. Le nystagmus de circumduction monoculair. Etudeclinique, oculographique et electro-myographiqued'un cas, dans la sclerose en plaques Rev Neurol(Paris) 1979;135:51-7.

4Findley LJ, Gresty MA, Halmagyi GM. A novel method

of recording arm movements: A survey of commonabnormalities. Arch Neurol 1980;38:38-42.

Hewlett Packard. Digital Signal Analyser User's Guide.HP 05420-900 32, July, 1978.

6Chokroverty S, Barron KD. Palatal myoclonus andrhythmic ocular movements: A polygraphic study.Neurology (Minneap) 1969;19:975-82.

'Nathanson M. Palatal myoclonus. Further clinical andPathophysiological observations. Arch NeurolPsychiatry 1956;75:285-96.

'Gresty MA, Halmagyi GM. Abnormal Head Move-ments. J Neurol Neurosurg Psychiatry 1979;42:705-14.

9 Ordenstein L. Sur la paralysie agitante et la sclkrose enplaques generalisee. Paris: Delahayet, 1868, 32 (ref toM Charcot).

Bender MB, Nathanson M, Gordon GG. Myoclonus ofmuscles of the eye, face and throat. Arch NeurolPsychiatry 1952;67:44-58.

Nashold BS, Slaughter DG, Gills JP. Ocular reactions inman from deep cerebellar stimulation and lesions.Arch Ophthalmol 1969;81:538-43.

"2Kornhuber HH. Motor functions of cerebellum andbasal ganglia: the cerebellocortical saccadic (ballistic)clock, the cerebellonuclear hold regulator, and thebasal ganglia ramp (voluntary speed smooth move-ment) generator. Kybernetk 1971;8:157-62.

13 Nashold BS, Gills JP. Ocular signs from brain stimula-tion and lesions. Arch Ophthalmol. 1967;77:609-18.

"Ron S, Robinson DA. Eye movements evoked by cere-bellar stimulation in the alert monkey. J Neurophysiol1973;36:1004-22.

"Bender MB. Brain control of conjugate horizontal andvertical eye movements. A survey of the structural andfunctional correlates. Brain 1980;103:23-69.

16 Llinas RR, Yarom Y. Electrophysiological properties ofmammalian inferior olivary cells in vitro. In: CourvilleJ, de Montigny C, Lemarre Y, eds. The Olivary Nuc-leus. Anatomy and Physiology. New York: RavenPress, 1980:379-88.

439

Protected by copyright.

on 6 August 2018 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.45.5.431 on 1 May 1982. D

ownloaded from