Journal of Neuro-Ophthalmology, December 1986, Volume 6, Issue 4

The cerebral ocular pursuit pathways. A clinicoradiological study of small-field optokinetic nystagmus.

Optokinetic nystagmus was tested in patients with intracerebral tumors restricted to one hemisphere, by using a simple, hand-held drum and visual evaluation of symmetry of horizontal response. Computed tomographic mapping revealed that lesions associated with asymmetric optokinetic nystagmus involved the posterior hemisphere. The critical region appeared to extend from the posterior splenium and the anteromedial occipital lobe to the posterior internal capsule. It seemed to course below the posterior horn of the ventricular system, then turn anteriorly, and finally medially, above the temporal horn. It is argued that the posterior extremity of this C-shaped region contains the ipsilateral and contralateral visual inputs to a parietal cortex control over ipsilateral smooth eye movements, and that the anterior extremity contains the efferent motor command pathways destined to the brain stem.

<c, 1986 Raven Press, New York The Cerebral Ocular Pursuit Pathways A Clinicoradiological Study of Small-Field Optokinetic Nystagmus Lena Kjallman, M.D., and Lars Frisen, M.D., Ph.D. Optokinetic nvstagmus was tested in patients with in­tracerebral tumors restricted to one hemisphere, by using a simple, hand-held drum and visual evaluation of svmmetry of horizontal resp(lnse. Computed tomo­graphic mapping revealed that lesions associated with asvmmetric (lptokinetic nvstagmus involved the poste­rior hemisphere. The critical region appeared to extend from the posterior splenium and the anteromedial oc­cipital lobe to the posterior internal capsule. It seemed to course below the posterior horn of the ventricular system, then turn anteriorly, and finally medially, above the temporal horn. It is argued that the posterior ex­tremity of this C-shaped region contains the ipsilateral and contralateral visual inputs to a parietal cortex con­trol over ipsilateral smooth eye movements, and that the anterior extremity contains the efferent motor com­mand pathways destined to the brain stem. Key Words: Intracerebral tumors-Optokinetic nys­tagmus. From the Department of Ophthalmology, University of G6teborg, Sweden. .. Address correspondence and reprint requl'sts to L. Frisl'n, Ogonkliniken, Sahlgren's Hospital, S-413 45 G6teborg, Sweden. 209 Optokinetic nystagmus (OKN) is a succession of involuntary eye movements induced by a moving series of visual stimuli, with a slow eye movement in the direction of the target, alternating with a re­setting fast movement. The normal symmetry of response for alternating directions of target move­ment may be lost with lesions of the brain and the brainstem. Although several studies have at­tempted to analyze the topographical relation­ships, particularly at the hemisphere level (1-3, and others), these remain poorly defined because of the considerable size of lesions in most of the reported cases and the relatively crude mapping techniques (angiography, surgical findings, or gross postmortem examinations). The modern tool of computed tomography (CT) offers not only a much better definition of lesions, but also a spatial framework that facilitates a search for any cores of injury that may be shared by subjects with any given symptom or sign. By using this approach, Kampf (4) found that OKN asymmetry frequently occurred with lesions involving the deep dorsal parietotemporal white matter. A more precise def­inition can be obtained by using a finer CT grid size, by excluding lesions with large dislocation effects, and by taking head size and shape into ac­count. Inclusion of cases with symmetrical OKN should enhance delimitation further. These tech­niques allowed us to define an envelope that in­cluded a portion of each lesion associated with asymmetric OKN and excluded most others. Disturbances of OKN also may occur with le­sions of the brainstem (5,6). This subject is outside the scope of the present investigation, however. CASES AND METHODS Selection of Cases Review of files at the Neuro-Ophthalmology Service at Sahlgren's University Hospital allowed 2/0 L. KIALLMAN AND L. FRISEN identification of 200 Cllnsecutive patients with a solitary intrahemispl1l'ric primary brain tumor, or ,1 single metastasis, who had been l'xamll1ed be­fl) rL' trt'atnwnt. Only intracerebral tumors were se­1L, l'tl'd because of their CT nll1spicuity and very IMgl' vari,ltion in location. Subjl.'cts who had spon­taI1l..' ous nystagmus, gazl' p,Hesis, or extraocular muscle palsy were l'xcluded. Optokinetic Testing OKN was tested with a hand-held motorized drum, which allowed instantaneous change of di­rection of rotation. Drum diameter was 180 mm. There were 14 black stripes separated by white stripes of equal width. With a testing distance of about 0.3 m, the rate of rotation corresponded to an angular velocity of 35 degrees per second. These parameters were found optimal in informal pilot studies. The patient's response was eva~u­ated by eye solely with regard to symmetry tor horizontally alternating directions of stripe move­ment. Any type of right-left asymmetry (e.g., in rhythm or amplitude of eye movements) was taken as abnormal, irrespective of its magnitude. Laterality was designated as the direction of drum movement (right or left, as seen by the patient) that was associated with the poorer response. Computed Tomography: Technique and Data Analysis All patients were examined in an EMI CT 1010 head scanner, using 10-mm slices, without overlap, parallel to the orbitomeatal plane. Iso­paque Cerebral (80 ml; Nyegaard, Oslo, Norway) was given intravenously for tumor enhancement. Polaroid copies of the enhanced scans were evaluated first with regard to brain size, because too large a variability would cause uncertainty in the subsequent pooling of data. It was decided to retain only cases where sagittal and transveral di­ameters in the section just above the suprasellar cistern fell within 1 SO of the adult average (185 :!: 14 mm, and 142 :!: 10 mm, respectively; data ob­tained from a random sample of 57 individuals and corrected for magnification). For each remaining patient, all CT sections containing an enhancing lesion or surrounding subnormal attenuation ("edema"), or both, were photographed at standard magnification. The negatives then were projected at standard magni­fication on transparent sheets at natural scale. Here, the outlines of the tumor and any sur- I Oill NCllrtl-tll'htllllllllol, Vtll. 6, No.4. 1981> 65 4 32 1o -1 -2 FIG. 1. Definition of CT planes. Nominal thickness 10 mm. rounding edema were traced by hand. If the max­imum section area of the tumor exceeded 15 cm2, it was considered excessively large and was ex­cluded from further analvsis. Otherwise, the mid­sagittal skull diameter w~s constructed from bony markers, and the intersections with the skull by its midpoint perpendicular were marked. Three-Dimensional Reconstruction and Pooling of Data A set of square-ruled (10 x 10 mm) transparent masters was prepared, one for each tomographic leveL numbered wnsecutively from the base of the skull (Fig. 1). For each master, the corre­sponding transparency was retrieved from each patient. The transparenc~"s diameters (as defined above) were aligned with thl)se L)f the master, and a mark was made in each master square that was covered to SLY, l)r more bv tumor. Different s~'mbols were used for patients with symmetric and asvmmetric OKN. RESULTS Forty-three patients (26 men and 17 women) fulfilled the inclusion criteria. Ages ranged from 16 to 71 years (median 57). Eighteen (42'7c) had asymmetric OKN, with the poorer response ob­tained on rotating the drum towards the side of the lesion. Inspection of the pooled observations revealed no consistent differences between left and right hemisphere observations. Therefore, all left-side observations were mirrored to the right side. Figure 2 summarizes the observations of sym- CEREBRAL OCULAf~ PURSUIT PATHWAYS 211 t ••• - - •. t . L.... .---1 .. 6 ··,'s"·" ",'·l'J'}'m I I '1...lJ • ~ 1 ) i 1 • , 1 ' I " -1 f')"ll -, .- t '. 'Wl '•. 1 .... + T . 4 ••• • .. Jl , " : j 1j r::~j j ~1 ) ) I +, I 1 , 1 1 1 1~ .~ I, • t t < ) 1 ) ) ..... , ,. t • 1 1 l .. . ",', 'I 'j , ' ., ,~,:,, , ,: • 1 I :' , , '~' I I / 1 , • 't I I .' I , .' I , , I , I , ',' • ' , ,. ·1 • FIG. 2. Distribution of lesions after mirroring left-hand observations to the right side. Anterior is up. Bold numerals identify craniocaudal level (cf. Fig. 1) Each square engaged by tumor contains one or two numerals: those in the lower right corner repre­sent the number of lesions asso­ciated with asymmetric OKN. and those in the opposite corner repre­sent lesions with normal OKN. Squares with a bold outline meet minimum criteria for core definition (cf. text); a bold diagonal signals preponderance of OKN asymmetry. Circles identify average endpoints of sagittal and transversal brain diam­eters. metric and asymmetric OKN responses for each tomographic level. Lesions causing asymmetric OKN cluster in levels 1 through 3 (d. Fig. 1) and involve the posterior half of the hemisphere. Scrutiny of Fig. 2 shows that the cluster repre­senting asymmetric OKN has a core defined by a preponderance of observations of OKN asym­metry, and a fringe with mixed symmetric and asymmetric observations. The cluster is sur­rounded more or less completely by symbols for symmetry. The occurrence of mixed observations may appear paradoxical, but is expected on the grounds of variations in brain size and tumor dis­location effects. Therefore, optimum definition of the core requires a weighting rule. We decided ar­bitrarily that each square containing three or more observations of the same OKN sign, or at least 75% same-sign observations out of four or more, should count as a criterion observation. Criterion squares have been given a bold outline in Fig. 2, with the addition of a bold diagonal to squares representing asymmetric OKN. Figure 3 is an oblique three-dimensional reconstruction based on the weighted data for asymmetric OKN, taking tomographic slice thickness into account. Translation of these findings into anatomical terms requires caution because of variations in brain size and shape, varying dislocation effects of the lesions, and the relatively coarse sampling grid. The following is a cautious attempt to de­scribe the extent of the roughly C-shaped region where a lesion is expected to cause OKN asym­metry. The posterior extremity emerges from the splenium and the anteromedial occipital lobe, and continues laterally, below the posterior horn of the ventricular system, toward the lower convexity of the parietal lobe. From here it makes a medial turn above the root of the temporal horn and then turns toward the vicinity of the posterior limb of the in­ternal capsule. The question whether the "edema" sur­rounding a tumor might affect OKN was an­swered in the negative after identifying several cases where the tumor itself was located outside the criterion region, but peritumoral edema ex­tended into the region. All these cases had sym- , eli" Neuro-ophthalmol, Vol. 6. No.4, 1986 III L. KIALLMAN AND L. FRISEN FIG. 3. Three-dimensional represen­tation of minimum criteria locations for asymmetric OKN (cf. squares with bold diagonals in Fig. 2), with mirroring to the left hemisphere. Hemispheres have been separated laterally for clarity. Lowest tomo­graphic plane is level 1 (cf. Fig..1). Curved lines represent average In­side contours of skull (cf. circles in Fig. 2). Stripes represent exposed ventricular surfaces. metrical OKN. Hence, asymmetry of OKN seems to occur only with damage that is more severe than that caused by edema. DISCUSSION The first clinical observation of asymmetry of OKN was reported by Barany in 1921 (7). Similar observations have been reported by many other investigators, but topographical interpretation has long been controversial. Although Cords's (1) hy­pothetical explanation involving a parietal cortex optomotor center and a corticofugal pathway re­ceived support in early clinical studies (2, 8, and others), many later investigators have concluded that the localizing value of OKN asymmetry is lim­ited to laterality (9,10). The latter view is explicable in terms of lesion size: studies of cases with large lesions easily obscure the possibility that any asymmetry of OKN can be attributable to a minor part of the lesion. OKN is composed of a slow pursuit eye move­ment and a resetting saccade. Impairment of ei­ther component for one direction of target move­ment causes lateral asymmetry of OKN. In prin­ciple, the hemisphere component of the saccadic control system can be injured anywhere between the cortex and the brainstem. The actual location of the cortical saccade control has long been elu­sive, but recent studies employing positron emis­sion tomography (11) have lent additional support to the classical emphasis on the frontal eye fields (12). Paradoxically, frontal lesions rarely are asso­ciated with asymmetry of OKN. This appears to be due to plasticity of cortical saccadic control: OKN I elin NClIro-"l'hlllllllll"l. V"I. b. No.4. 198b asymmetry occurs only evanescently with acute frontal lobe lesions, until the contralateral frontal eve field takes over (13,14). Hence, in cases with n"onacute lesions, asymmetry of OKN is attribut­able to impaired slow eye movement control (15). Actually, there are two different systems for the control of pursuit eye movements. One seems to depend on information coming from the periph­eral visual field, over the visual accessory system, and is not involved with hemispheric lesions as discussed here. The other system, which normally overrides the brainstem pursuit system, depends on foveal visual input and normal hemisphere function. This is the system tested by a small OKN drum. Testing of the brainstem system requires a large-field display and laboratory resources (16). The location of cortical control over slow eye movements is not known precisely in man, al­though clinical observations implicate parietal cortex (8,17,18). Experimental studies in the monkey implicate the inferior parietal lobule, prestriate cortex, and possibly the middle tem­poral visual area (19,20). The required input from visual cortex appears to be both ipsilateral and contralateral, because unilateral occipital lesions do not affect ipsilateral pursuit or OKN (14,21). The actual pathways mediating visual information are not known, however. Likewise, the pathways ior efferent slow eye movement commands are poorly known, with little specific support for sug­gestions like Flechsig's optomotor pathway, in­ternal corticotectal or corticopontine tracts, or the internal sagittal stratum (15,22,23). They seem to leave the hemisphere in or close by the posterior limb of the internal capsule (8,24). CEREBRAL OCULAR PURSUIT PATHWAYS 2/3 FIG. 4. Proposed course of right hemisphere pathways involved in cerebral ocular pursuit control to the right; schematic representation seen from above. Anterior is up. Presumed cortical control area (bold outline) receives both contralateral (transcal­losal) and ipsilateral visual input (black arrows) over pathways running below posterior horn of ventricular system. Efferent motor commands project above tem­poral horn towards posterior limb of internal capsule (dotted area). Our CT analysis (Figs. 2-4) can offer no more than a coarse picture of the spatial distribution of the pathways just discussed. It is possible that better detail might be obtained by using a finer CT grid, and by developing techniques for accommo­dating the full spectrum of brain sizes and config­urations. The potential gain has to be weighed against the quite considerable size of most lesions at the time of clinical presentation, however. It is also possible that quantitative analysis of oculo­graphic records may reveal subtler degrees of ab­normality than mere visual inspection. Neverthe­less, it appears that the simple techniques used here are powerful enough to merit use in routine clinical work. On the basis of our CT analysis and the investi­gations related above, we propose that the pathways involved may have the following courses (Fig. 4). Visual premotor input to the right pursuit cortical control arrives over both contralat­eral transcallosal and ipsilateral fibers. These seem to run below the posterior ventricular horn en route to the inferior parietal lobule. From here, the efferent motor command pathways run anteriorly and turn medially above the root of the temporal horn, and then turn towards the posterior limb of the internal capsule. A lesion anywhere in this system will cause asymmetry of OKN, with the poorer response obtained on target movement to the side of the lesion. Such a lesion mayor may not also involve the optic radiations. A homon­ymous visual field defect does not by itself cause OKN asymmetry, because the cortical pursuit con­trol, as tested by a simple manual OKN drum, ap­pears to work equally well with transcallosal input alone, as with ipsilateral or bilateral visual input. The most likely locations of a lesion causing both asymmetric OKN and homonymous visual field defects are the upper posterior temporal lobe and the lower parietal lobe. However, even for these locations, it is conceivable that the pursuit com­mand pathways may be affected in isolation. Acknowledgment: Hans Kjallman made the three-di­mensional reconstruction. REFERENCES 1. Cords R. Optisch-motorisches Feld und optisch-motorische Bahn. Ein Beitrag zur Physiologie und Pathologie der Rin­deninnervation der Augenmuskeln. Arch Ophthal11l0/ 1926;117:58-113. 2. Kestenbaum A. Zur topischen Diagnostik der Hemian­opsie. Z AlIgmhelikd 1932;76:241-77. 3. Enoksson P. Optokinetic nystagmus in brain lesions. Acta Ophthalllwi 1956;34:163-84. 4. Kampf D. The significance of optokinetic nystagmus asvm­metry in hemispheric lesions. Nellro-ophtI1ll/11I0/oglf 1986;6: 61-4. 5. Tos M. Adser I, Rosborg }. Horizontal optokinetic nys­tagmus in cerebral diseases. Acta Nellro/ Scand 1972;48:607­20. 6. Baloh RW, Honrubia V, Sills A. Eve tracking and optoki­netic nystagmus-results of quantitative testing in patients with well-defined nervous system lesions. Ann Oltl/ RIll/wi Lan/Ilgo/1977;86:1-7. . 7. Baranv R. Zur Klinik und Theorie des Eisenbahnnvs-tagmu~. Acta Ottl/an/ngt]/ 1921;3:260-5. . R. Fox }e. Holmes G. Optic nystagmus and its value in the location of cerebral lesions. Bra/n 1926;49:333-71. 9. Davidofi RA, Atkin A. Anderson PI, Bender MB. Optoki­netic nystagmus and cerebral disease. Clinical and patho­logical study. Arch NellroI1966;14:73-81. 10. Takemori S, Ishikawa M, Yamada S. Cerebral control of eve movements. ORL 1981;43:262-73. . 11. Fox PT, Fox }M, Raichle ME, BurdI' RM. The role of cere­bral cortex in the generation of voluntary saccades: a posi­tron emission tomographic study. I Nellrophl/sio/ 1985;54: 34R-69. 12. Bruce Cj, Goldberg EM. Phvsiology of the frontal eye fields. Trends Nellro/ Sci 19R4;7:436-41. 13. Hoyt WF, Darofi RB. Supranuclear disorders of ocular con­trol systems in man. Clinical, anatomical, and physiolog­ical correlations-1969. In: Bach-y-Rita P, Collins Ce. eds. The CO/Itro/ 0' el/e lIIovelllents. New York: Academic Press, 1971:175-235.. 14. Sharpe }A, La AW, Rabinovitch HE. Control of the saccadic and smooth pursuit systems after cerebral hemidecortica­tion. Bra/II 1979;102:387-403. I Clin Ncuro-0l'lJllJalmol. Vol. 6, No.4, 1986 21.+ L. KIALLMAN AND L. FRISEN 15. B.ll,.h I~W. h'" I{[), I {onrubi.l V. Oplokinl'lic nysl'lgmus •1I1d p.Hil'l,lllllb,' Il'sions. /\111/ Nom.1 IlJHII;7:2hlJ-7h. Ih. Il,ll"h InV, y"" [{D, IllInrubi,l V. Clinic,1i dbnorm,llitil's lIf ,'pl"l..inl'lic nl'st,lgmus. In: 1."I1I1l'rstr.lnd C, h'" DS, Kl'IIl'r n., "ds. 1'1I11<'1i"1I111 /"I.'i" ,., ,.,'11/111' 1I1",t/,II/ "i,,,nl,'h. t hford: 1\'rg'1Il1111l I'n'ss, I'IH2:.1II· 211. . 17. Smith II., ClIg,ln DC. llptlll..illl'lic nvst'lgmu.-: ,I tl'st tor pa­ril'I, llllIb,' I"silln.s. :\/11 J I l/'"'"IIIIII,,1 14<;'J;"H: IH7 -'1.1. Ill. Smith Jl.. I l/",.kllldi.. 111/.,111.'\11111.-. Ih 11,1' ill '''/,Iu/I /1I'1If1I-1'I,II· (//11/I11"I,'.'\"·,t/ "ill.'\I/",i". Springlil'ld: CC Thonh", l'Ih3. 1'1. Ll'nch JC Md.olr,'nIW. t lplol..il1l'lic nysl,'gllllis dl'ficits fol­I" wing p.Hil'l,"occipil.ll clIrlex "'sions in mllnkl'ys. Lx/' Nn/lll I{t'" I'IH3;"'J: 12<;-.111. 20. N"WSlllll" WT, Wurtz RH, Durst"'er MI{, Mikami A. Del· I ClIII Nt'lIrt""l'hl/ltlll11ol, Vol. b, No . .J, 191i6 icits in visual motion processing following ibotenic acid le­sions of the middle temporal area of the macaque monkey. J Nt'l/rosei 1985;5:825-40. 21. Sharpe JA. Cerebral ocular motor deficits. In: Lennerstrand G, Ze,' OS, Keller EL, eds. Functional basis of ocular motility di'''rdcrs. Oxford: Pergamon Press, 1982:479-88. 22. Leigh RJ, Zee OS. Tile ncuroioxy of cye 1I/0VClI/ellts. Philadel­phia: Davis, 1983:79-80. 23. Sharpe JA, Deck JHN Destruction of the internal sagittal stratum and normal smooth pursuit. Ann Neural 1978;4: "7l-h. 24. Hirose C, Kosoegawa H, Saeki M, et al. The syndrome of poskrior thalamic hemorrhage. Ncurology 1985;35:998­1U1I2.