Journal ofNeurology, Neurosurgery, and Psychiatry 1995;58:91-94 91 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.58.1.91 on 1 January 1995. Downloaded from SHORT REPORT

Horizontal disorders after posterior vermis infarctions

K Vahedi, S Rivaud, P Amarenco, C Pierrot-Deseilligny

Abstract of this study was to analyse quantitatively sac- The horizontal saccade, smooth pursuit, cade accuracy and smooth pursuit velocity in and vestibulo-ocular reflex gains were patients with a recent and pure cerebellar recorded in 19 patients with cerebellar infarction, documented by MRI. infarction documented with MRI, and in a group of control subjects. Bilateral sac- cade hypometria and a decrease in ipsi- Patients and methods lateral smooth pursuit gain were found Nineteen patients (mean (SD) age 56 (15) only in patients with a lesion affecting years) with a cerebellar infarction were stud- the posterior vermis. These results in ied. Patients with a coexisting humans support experimental findings infarction or any clinical sign or radiological suggesting that the posterior vermis con- evidence of oedematous brainstem compres- trols both saccade accuracy and smooth sion were not included. The location of the pursuit velocity. infarction was defined from MRI axial sec- tions and previously published templates (3 Neurol Neurosurg Psychiatry 1995;58:91-94) based on pathological studies of 64 cerebella.8 10 Coronal and sagittal sections were also cor- related with Courchesne's templates." 12 Two Keywords: ; eye movements patient groups were constituted according to presence and absence ofvermal involvement. In the monkey, two cerebellar regions, the and the posterior vermis, are known GROUP 1 to be involved in smooth pursuit. Bilateral This group comprised nine patients (mean ablation of the flocculi and portions of the (SD) age 65 (11) years) with an infarction paraflocculi results in a reduction of smooth involving the cerebellar vermis (figure). The http://jnnp.bmj.com/ pursuit velocity to 65% of normal values, vascular territory involved was that of the which decreases to 30-40% with the addi- posterior inferior cerebellar artery, on the tional ablation of the posterior vermis (lobuli right side in six cases and on the left side in IV to VIII).I Neural activity related to smooth three (one case associated with an ipsilateral pursuit has been recorded in these two struc- infarct involving the superior cerebellar artery tures.2 Furthermore, the posterior vermis and territory). In the cerebellar vermis, the infarc-

underlying fastigial nuclei are involved in the tions affected variably lobuli VI to X (the on September 24, 2021 by guest. Protected copyright. control of saccade accuracy, as ablation of clivus, tuber, pyramis, uvula, and nodulus), these structures in the monkey results in sac- corresponding to the posterior vermis. Lobuli cade dysmetria.3 4 Neural activity related to VIII and IX were damaged in all cases and saccades has been recorded in vermal lobuli lobuli VI and VII in four cases. The inferior Service de Neurologie, VI and VII in the monkey.2 Anatomical data part of the ipsilateral H6pital Saint- suggest that the oculomotor vermis projects was also damaged, including variably the ton- Antoine, Paris, France predominantly to the ipsilateral mediocaudal sil, the lobuli biventer, gracilis, semilunaris K Vahedi fastigial .5 inferior, and the inferior part of the lobulus P Amarenco The ocular motor signs resulting from semilunaris superior. The flocculus and the Service de Neurologie lesions affecting specific cerebellar regions dentate nuclei were not affected by any of the and Unite INSERM have not yet been precisely defined in infarctions. Brain MRI did not show recog- 289, Hopital de la Salpftrikre, Paris, humans. Electro-oculographic studies of nisable fastigial nuclei. Eye movements were France patients with cerebellar lesions have often recorded on average 35 (SD 42) days after S Rivand involved degenerative, tumorous, or demyeli- onset of the stroke. C Pierrot-Deseilligny nating processes, usually not confined to the Correspondence to: Dr K Vahedi, Service de cerebellum. Electro-oculographic studies of GROUP 2 Neurologie, H6pital Saint- patients with focal and pure cerebellar lesions This group comprised 10 patients (mean Antoine, 184 rue de Faubourg Saint-Antoine, have rarely been performed.67 Because infarc- (SD) age = 47 (12) years) with a cerebellar 75012 Paris, France. tion topography has been described in recent infarction sparing the cerebellar vermis (fig- Received 3 June 1994 pathological and radiological studies, MRI ure). The vascular territory was that of the and in revised form may be used 1 September 1994. for the radioanatomical correla- right posterior inferior cerebellar artery in Accepted 9 September 1994 tion of such lesions.8 9 Therefore, the purpose three cases and that of the superior cerebellar 92 Vahedi, Rivaud, Amarenco, Pierrot-Deseilligny

Drawing ofthe cerebellar J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.58.1.91 on 1 January 1995. Downloaded from areas involved by the infarcts as they appear in MRI on horizontal sections, independent ofthe infarct side. (Sectionsfrom the rostral to the caudal cerebellum: (1) and (2) sections through the upper _@ . pons, (3) section through the midpons, (4) section 1 2 through the upper medulla.) A = group 1, B = group 2, L = left, R = right. The posterior vermis is represented by the hatched area.

3 4 A B

artery in seven cases (two on the left side and the subject to follow a light emitting diode five on the right side). The infarction area moving sinusoidally, with an eccentricity of affected variably the tonsil, the lobuli biven- ±200, peak velocity of 230/s (frequency = ter, gracilis, and the semilunaris inferior, 0-28 Hz), and 45°/s (frequency = 0O37 Hz). superior, simplex, and anterior. The dentate The smooth pursuit gain (peak eye velocity nucleus was damaged in two patients. The over peak target velocity) was determined for flocculus was spared. Eye movements were each target velocity and each horizontal direc- recorded on average 26 (SD 19) days after tion by averaging the gain existing in 10 con- onset of the stroke. secutive cycles. The vestibulo-ocular reflex (VOR) was CONTROL GROUP tested by rotating the subjects sinusoidally in This group included nine normal subjects darkness while fixating an imagined earth- (mean (SD) age = 57 (8) years). fixed target, with an amplitude of ±200 (fre- quency = 0-3 Hz, peak velocity = 35°/s). The EYE MOVEMENT RECORDINGS peak eye velocity over peak rotation velocity Eye movements were recorded by direct cur- (VOR gain) was determined for both horizon- rent electro-oculography in darkness, with the tal directions by averaging the results of 10 immobilised. Three types of horizontal consecutive cycles. eye movements were studied in the patient and control groups. STATISTICAL METHODS For reflexive visually guided saccades, the The saccadic, smooth pursuit, and VOR subject was instructed to fixate a central point gains in each patient group were compared and then to look as quickly as possible at a with those of the control group by analysis of http://jnnp.bmj.com/ lateral target suddenly occurring 250 left or variance, and multiple comparisons were right of the central point with unpredictable made with the Newman-Keuls procedure. timing and direction. The saccadic gain (ratio The comparison between the ipsilateral and of primary saccade amplitude to target eccen- contralateral values was made by Student's t tricity) was determined for four types of sac- test. The gain for each control subject and cades: from the central position to the side of each type of eye movement was obtained by the lesion (ipsilateral centrifugal saccade) or averaging the rightward and leftward values, on September 24, 2021 by guest. Protected copyright. the opposite side (contralateral centrifugal as no significant difference was found between saccade), and from the side of the lesion the two types of values in these subjects. (ipsilateral centripetal saccade) or the oppo- site side (contralateral centripetal saccade) to the central position. The results were aver- Results aged for each type of saccade with 15 consec- Table 1 gives the mean results. In group 1 the utive saccades in each subject. saccadic gain was decreased in both centrifu- Smooth pursuit was tested by instructing gal directions compared with that of controls

Eye movement results Saccade gain (mean (SD)) Smooth pursuit gain (mean (SD)) VOR gain (mean (SD)) Centrifugal Centripetal 23°/s 45°ls 35°/s Ips Cont Ips Cont Ips Cont Ips Cont Ips Cont Group 1 0-82**t 083*t 0-89 0-89 068*t 0-78 043**tt 0-67 0-83 0-92 (n = 9) (0 08) (0 08) (0.12) (0-10) (0 24) (0-19) (0 24) (0 27) (0-15) (0 27) Group 2 0-91 0 93 0 95 0-96 0-93 0-92 0-86 0-83 0 93 0 90 (n = 10) (0 04) (0 06) (0 04) (0 05) (0-04) (0 08) (0-15) (0-15) (0-19) (0-12) Controls 0 95 0 97 0-95 0-85 0-83 (n = 9) (0 03) (0 03) (0-07) (0 06) (0-11) *p < 0-01; **p < 0-001 v controls; tp < 0-01; ttp < 0 00 group 1 v 2. Ips = ipsilateral; Cont = contralateral. Horizontal eye movement disorders afterposterior vermis infarctions 93 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.58.1.91 on 1 January 1995. Downloaded from (p < 0001 for ipsilateral saccades and p < Therefore, in the case of midline cerebellar 0 01 for contralateral saccades) or to that of lesions, it may be that hypermetria results group 2 (p < 0 01 in both directions). The mainly from damage to the fastigial nuclei. mean undershooting (17-18%) was similar Bilateral hypometria of mainly centrifugal for both centrifugal directions. Centripetal saccades was the pattern of dysmetria in our saccades were only slightly hypometric in patients with damage to the posterior vermis. group 1. In group 2, the centrifugal and cen- In most of them, the strictly unilateral nature tripetal saccade gains were similar to those of of the lesion could not be ascertained with controls. MRI, as the vermis is located on the midline. In group 1, the smooth pursuit gain was Furthermore, infarctions affecting this struc- decreased, slightly for the contralateral direc- ture often extend slightly beyond the tion but significantly for the ipsilateral direc- midline.8 Such hypometria of centrifugal sac- tion, compared with the values of the control cades has previously been found in experi- group (p < 0-01 at 230/s and p < 0001 at mental studies in the monkey, after bilateral 45°/s) or with those of group 2 (p < 0-01 at posterior vermal lesions.317 Although it was 23°/s and p < 0O001 at 45°/s). In group 1, the not possible in our patients to determine with smooth pursuit gain was variably decreased, MRI whether the fastigial nuclei were explaining the relatively large SDs in this affected by the infarctions, such damage group. It was not possible, however, to corre- seems unlikely as: (1) hypermetria was not late this gain to the variable extent of the found in these patients; (2) the vascular terri- lesions within the vermis (or in the paraver- tory of the fastigial nuclei is that of the supe- mal structures) because the resolution of rior cerebellar artery, and not that of the MRI is insufficient to quantify the volume of posterior inferior cerebellar artery,89 involved such small lesions. A significant asymmetry in group 1. Therefore, it may be hypothesised between the ipsilateral and contralateral gains that hypometria of centrifugal saccades in our existed in group 1 (p < 0-05). In group 2, the patients was due to damage to the posterior smooth pursuit gain was similar to that of vermis and not to the fastigial nuclei. controls. For each patient group and each direction, SMOOTH PURSUIT the VOR gain was not significantly different Because the brainstem, cerebellar peduncles, from that of the control group. No abnormal and flocculus were spared in our patients, the eye movements, such as square wave jerks or smooth pursuit deficit found in group 1 con- nystagmus, were found in either group. firms that the posterior cerebellar vermis is involved in the control of this eye movement in humans.8 The smooth pursuit gain was Discussion decreased bilaterally, but only significantly so Saccade hypometria and a decrease in the for ipsilateral movements. This is consistent smooth pursuit gain existed in group 1 but with electrophysiological studies in the mon- not in group 2. The main difference between key, in which it has been reported that 60% the two groups was damage to the posterior of the posterior vermis cells show a response vermis in group 1. Furthermore, in all during ipsilateral and 40% during contralat- the infarction was purely cerebellar, eral smooth patients, pursuit.' http://jnnp.bmj.com/ without brainstem compression. The floccu- In conclusion, these results confirm that lus was never damaged. Therefore, eye move- the posterior vermis is involved in eye move- ment abnormalities observed in group 1 ment control in humans. Although small ver- probably resulted from damage to the poste- mal infarctions are rare, further human rior vermis. studies with such lesions and other eye move- It should be noted that, in the posterior ment stimuli are required to determine the vermis, lobuli VI to IX were variably dam- specific role of the different vermal lobuli and

aged in group 1, with only lobuli VIII and IX the other cerebellar structures in this control. on September 24, 2021 by guest. Protected copyright. as areas of common damage. In monkeys, lobulus VIII may be involved in smooth pur- 1 Keller EL. Cerebellar involvement in smooth pursuit eye suit, but lobuli VI and VII seem to be the movement generation: flocculus and vermis. In: main lobuli involved in control of eye move- Kennard C, Clifford Rose F, eds. Physiological aspects of clinical neuro-ophthalmology. London: Chapman and ment.2 Even though lobuli VI and VII were Hall, 1988:341-55. not damaged in all the patients of group 1, 2 Sato H, Noda H. Posterior vermal Purkinje cells in macaques responding during saccades, smooth pursuit, their afferent tracts from the or chair rotation and/or optokinetic stimulation. Neurosci their efferent tracts to the fastigial nuclei513 Res 1992;12:583-95. 3 Ritchie L. Effects of cerebellar lesions on saccadic eye could, however, have been impaired by the movements. J Neurophysiol 1976;39: 1246-56. nearby lobulus VIII lesion. 4 Optician LM, Robinson DA. Cerebellar-dependent adap- tive control of primate saccadic system. Y Neurophysiol 1980;44: 1058-76. SACCADES 5 Yamada J, Noda H. Afferent and efferent connections of the oculomotor cerebellar vermis in the macaque mon- Hypermetria of centripetal saccades has been key. J Comp Neurol 1987;265:224-4 1. reported in the monkey after experimental 6 Ranalli PJ, Sharpe JA. Contrapulsion of saccades and ipsi- lateral : a unilateral disorder of the rostral cerebel- lesions affecting either the posterior vermis lum. Ann Neurol 1986;20:311-6. and the underlying fastigial nuclei4 or these 7 Pierrot-Deseilligny C, Amarenco P, Roullet E, Marteau nuclei been R. Vermal infarct with pursuit eye movement disorders. only.'4 Hypermetria has also J Neurol Neurosurg Psychiatry 1990;53:519-21. reported in patients with midline cerebellar 8 Amarenco P, Hauw JJ. Anatomie des arteres cere- belleuses. Rev Neurol 1989;145:267-76. lesions affecting the posterior vermis and 9 Amarenco P. The spectrum of cerebellar infarctions. probably also the fastigial nuclei.5 16 1991;41:973-9. 94 Vahedi, Rivaud, Amarenco, Pierrot-Deseilligny

10 Amarenco P, Hauw JJ, Caplan LR. Cerebellar infarctions. 14 Villis T, Hore J. Characteristics of saccadic dysmetria in J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.58.1.91 on 1 January 1995. Downloaded from In: Lechtenberg R, ed. Handbook of cerebellar diseases monkeys during reversible lesions of medial cerebellar New Jersey: Marcel Decker, 1993:251-90. nuclei. Neurophysiol 1981;46:828-38. 11 Courchesne E, Press GA, Murakami J, et al. The cerebel- 15 Selhorst JB, Stark L, Ochs AL, Hoyt WE. Disorders lum in sagittal plane-anatomic-MR correlation: I. The in cerebellar ocular motor control. I. Saccadic over- vermis. AJNR Am Neuroradiol 1989;10:659-65. shoot dysmetria: an oculographic control system, 12 Press GA, Murakami J, Courchesne E, et al. The cerebel- and clinico-anatomical analysis. Brain 1976;99: lum in sagittal plane-anatomic-MR correlation: II. The 497-508. cerebellar hemispheres. AJ7NR Am Neuroradiol 1989; 16 Buttner U, Straube A, Spuler A. Saccadic dysmetria and 10:666-76. "intact" smooth pursuit eye movements after bilateral 13 Thielert CD, Thier P. Patterns of projections from the nuclei lesions. J7 Neurol Neurosurg Psychiatry 1994;57: pontine nuclei and the nucleus reticularis tegmenti pon- 832-34. tis to the posterior vermis in the rhesus monkey: a study 17 Sato H, Noda H. Saccadic dysmetria induced by transient using retrograde tracers. Jf Comp Neurol 1993;337: functional decortication of the cerebellar vermis. Exp 113-26. Brain Res 1992;88:455-8. http://jnnp.bmj.com/ on September 24, 2021 by guest. Protected copyright.