OCULOMOTOR ORGANISATION Arris and Gale Lecture delivered at the Royal College of Surgeons of England on 15th December 1955 by Roger Warwick, B.Sc., M.D. University Professor of Anatomy, Guy's Hospital Medical School; Late Lecturer in Anatomy, University of Manchester IT IS FIRST my duty and pleasure to acknowledge the honour done me by election to this Arris and Gale Lectureship. As I believe is customary, the subject which I bring before you is fundamentally of academic interest, and since I am by profession an anatomist, I must treat extensively of structure. However, a purely topographical account of the motor innervation of ocular muscles would doubtless be to you, and certainly to me, a topic of limited appeal. I am therefore sure you will be relieved to learn that we shall be concerned with the functional significance of the structures we are to consider at every stage. Elucidation of the linkages between form and function provides, of course, the keenest pleasure in any biological enquiry, medical or otherwise; and it is this pleasure which is often the mainspring of academic research, as in my own. Nevertheless, you will discover in the observations here presented some points of practical significance in the semantics of clinical neurology, regarding which many here will be more able judges than I. Oculomotor organisation has for long attracted minds of contrasting interests, and even a rapid glance at the century of endeavour behind us today will show how the efforts of pure anatomists, experimenters, clinical observers, and even the purely speculative, have combined in the evolution of present conceptions. Stilling (1846) might be called the prime topographer of the cranial nerve nuclei. Certainly he first clearly revealed the sources of the third, fourth, and sixth nerves, amongst others, and his descriptions remained unchanged until 1881, when von Gudden, an experimentalist, demon- strated the chronic atrophy produced in nerve cells by division of their axons in young animals. His chief contributions, achieved by this method, were that the trochlear decussation is complete, that some oculomotor fibres cross within the midbrain, and that each of the two third nerve nuclei is divided longitudinally into dorsal and ventral columns. Apart, however, from the obvious deduction that the trochlear and abducent nuclei are the motor pools of the superior oblique and lateral rectus, his technique yielded no results concerning the central arrangement of the motor cells of the remaining ocular musculature. A few years later Edinger (1885) and Westphal (1887) described the small-celled nuclei which still commemorate them, and in 1889 Perlia combined these and his own topographic findings to produce his well- known diagram of the oculomotor nuclei (Fig. 1). He added another paired and small-celled group, the anteromedian nucleus, situated 36 OCULOMOTOR ORGANISATION Fig. 1. Topographical disposition of the oculomotor complex of nuclei according to Perlia (1889). Note inclusion of Darkschewitsch's nucleus as a radicular element, separateness of the Edinger-Westphal and anteromedian nuclei, and the inclusion of the nucleus centralis. cranially and in the midline, and caudal to this appeared his better known " central " nucleus. With both of these we shall be much concerned. He also favoured a more elaborate subdivision of the main, or lateral, oculomotor nucleus than von Gudden (1881). He included in what might henceforward be called the oculomotor complex not only the above elements but also the recently discovered nucleus of Darkschewitsch (1889) and a further median group of somewhat scattered neurones caudal to the central nucleus, which does not appear in his diagram, but is usually referred to as the " diffuse midline nucleus of Perlia." Prior to Perlia's topographical foundation, Adamuk (1870) and Hensen and Volckers (1878) had already found by midbrain stimulation, 37 R. WARWICK 38 OCULOMOTOR ORGANISATION mechanical and electrical, in the vicinity of dogs' oculomotor nuclei, that motor cells of the individual muscles innervated by the third nerve were not intermingled, but segregated into a cranio-caudal string of motor pools or " centres " (Fig. 2). Clinicopathological observers, such as Kahler and Pick (1881) and Starr (1888), were also deducing similar schemes in man, but these may well have been influenced by the experi- mental findings. Knies (1891) went much further in assigning functional significances to the various elements of Perlia's diagram (Fig. 2). Little weight should ever have been attached to this arbitrary proceeding for it was founded on little substantial evidence. Unfortunately Knies' purely speculative labelling of Perlia's central nucleus as a centre for convergence, a superficially attractive innovation, aroused immediate and lasting interest, despite the fact that other clinicians, such as Siemerling (1891), who described the clinical and pathological findings in numerous cases of ophthalmoplegia, could not be equally dogmatic. The next great impetus to study of the problem of muscle representation within the oculomotor hierarchy was provided by Nissl (1892, 1894) when he evolved his method of tracing neuronal connexions by observation of the disappearance of chromatin granules in response to axonal injury. Several experimenters at once applied the Nissl technique by cutting the nerve supplies of extra-ocular muscles and studying the distribution of affected nerve cells in the oculomotor nuclei. Among these Bernheimer (1897) and Bach (1899) were prominent. Their results were markedly INTRINSIC MUSCLES LEVATOR RECTUS MEDIALIS PALPEBRAE (IN CONJUGATE SUPERIORIS DEVIATION) RECTUS SUPERIOR MEDIALIS OBLIQUUS ~~~~~RECTUSCONVERGENCE) INFERIOR ~~~~(IN RECTUS INFERIO ONBUQUUS US- - -TROCHLEAP. -'-.- NUCLEUS Fig. 3. Oculomotor organisation according to Brouwer (1918). This is the diagram most frequently reproduced in text-books. Note dual motor pools for the medial rectus muscles. 39 4 R. WARWICK LEVATOR INTRINSIC MM. Fig. 4. Bernheimer's schema of the oculomotor nuclei. Note varying degrees of crossed innervation of the rectus medialis (R. int.), rectus inferior (R. inf.), and and obliquus inferior (0. inf.) opposed (Fig. 2). Perhaps because he used monkeys, perhaps owing to Brouwer's approval at a later date, Bernheimer's schema was preferred, and still forms the main factual basis of the conception of oculomotor organisation followed in most current accounts (Fig. 4). Although, as we shall see, Bach's views were perhaps nearer the truth, Brouwer (1918) accepted Bernheimer's pattern of motor pools for the extra-ocular muscles, but he modified it by banishing innervation of intiinsic ocular musculature from the central nucleus of Perlia (Fig. 3). His own comparative observations, eked out by data from a single case of ophthalmoplegia, led him to support Knies' designation of the central nucleus as a convergence centre. His emphasis on this has doubtless helped to engrave it in orthodox teaching, and although during the subsequent thirty-seven years further clinical, comparative, and experimental studies have continued to appear, these have had little effect upon authoritative texts. 40 OCULOMOTOR ORGANISATION Hadidian and Dunn (1937) employed Nissl's retrograde degeneration technique in goldfish, as did Abd-el-Malek (1938) in cats. Cats were also used in experiments of Szentagothai (1942, 1943) and Danis (1948) but their method was to observe ocular movements caused by stereotaxically controlled stimulation of points in the midbrain, a somewhat indirect procedure also selected by Bender and Weinstein (1943), who were the first workers since Bernheimer (1897) to experiment on monkeys in studying this problem. The results of these later investigations not only differed from the classical Bernheimer-Brouwer picture of oculomotor organisations but also amongst themselves. The origin of parasympathetic fibres for the innervation of the sphincter pupillae and ciliaris oculi was thought to be in the Edinger-Westphal nuclei by Bernheimer (1897) and Brouwer (1918), but this was denied by Bach (1899), Latumeten (1924) and others. Further matters of uncer- tainty exist, such as the proportion of oculomotor fibres which decussate, the identity of muscles supplied by crossed fibres, the role of certain close neighbours of the accepted oculomotor components, like the nucleus of Darkschewitsch, and so on. Moreover, as often occurs in brainstem topography, confusion has not infrequently arisen through independent description and labelling of the same nuclear group by several observers. Thus Perlia's diffuse nucleus of the midline is evidently the same entity as the nucleus dorso-centralis of Panegrossi (1898) and the caudal central Rostral Dtir g h~~~~~~~~~~~~Rgt Dorsal Right~orsa R. lateral aspect Dorsal Dorsal aspect Rostral section Reproduced by permission of the Journal of Anatomy. Fig. 5. Diagrams showing the topographic structure of the monkey's oculomotor nuclei. The small celled, or accessory, components are shown in black. Contrast with the functional schemes displayed in Figs. 2 to 4. 41 4-2 R. WARWICK nucleus of Tsuchida (1906), a fact not always recognised in topographic and other accounts (Warwick, 1953c). Perlia's pioneer work on the arrangement of the oculomotor nuclei has, of course, been considerably amplified and modified by authorities such as Tsuchida (1906), Frank (1921), Le Gros Clark (1926), Pearson (1944), Crosby, Henderson and Woodburne (1943), and Olszewski and Baxter (1954). Indeed, although