Superior Oblique Muscle Layers in Monkeys and Humans

Abstract

purpose. Rectus and the inferior oblique extraocular muscles (EOMs) consist of orbital layers (OLs), inserting on connective tissues, and global layers (GLs), inserting on the sclera. This study was performed to clarify the anatomic relationships of the corresponding layers of the superior oblique (SO) muscle.

methods. Two whole human and two monkey orbits were serially sectioned en bloc at 10-μm thickness in the coronal plane and stained for collagen with Masson’s trichrome and for elastin with van Gieson’s stain. The SO muscles of one human and one monkey were sectioned longitudinally. The structure of the SO muscle was examined by light microscopy, and muscle fibers in the OL and GL of selected sections were counted.

results. The deep SO muscle consisted of a central GL contiguous with the tendon, surrounded coaxially by a peripheral OL inserting on the SO sheath posterior to the trochlea. The maximum number of SO fibers was 14,400 to 19,200 in the human and 7,000 to 7,400 in the monkey. In the monkey, approximately 60% of total fibers were in the GL, and 40% in the OL. The SO sheath was in mechanical continuity with the superior rectus pulley.

conclusions. The primate SO has a substantial OL configured to contribute to positioning the superior rectus pulley in the coronal plane. Whereas the direction of application of the SO’s GL force is determined by the rigid trochlea, the SO’s OL influences the direction of application of rectus EOM forces. This insight extends the concept of active control of pulley positions to include a contribution from the SO muscle.

The oculorotary extraocular muscles (EOMs), but not the lid-elevating levator palpebrae superioris, consist of two layers. ¹ The global layer (GL), containing in humans a maximum of 10,000 to 15,000 fibers in the midlength of the EOM, is located adjacent to the globe in rectus EOMs. ² In the rectus EOMs, the GL anteriorly becomes contiguous with the terminal tendon and inserts on the sclera. ³ The GL of the inferior oblique (IO) muscle inserts directly on the sclera without a tendon. ⁴ The orbital layer (OL) of each rectus EOM contains 40% to 60% of all the EOM’s fibers, ² and the OL of the IO is also substantial. ⁴ The OLs of the rectus and IO EOMs do not insert on the eyeball, but instead insert on connective tissue pulleys. The OL is located on the orbital surface of the rectus EOM, sometimes forming a C-shaped configuration, and reportedly constitutes a roughly concentric outer layer of the oblique EOMs. ¹ The OL of the IO inserts on connective tissues including the pulleys of the inferior (IR), lateral rectus (LR), and the IO. ⁴  

Each rectus pulley is a relatively soft structure, consisting of an encircling sleeve and ring in Tenon’s fascia located near the globe equator ^{3 5 6 7 8} and coupled to the orbital wall, adjacent EOMs, and equatorial Tenon’s fascia by bands containing connective tissue and smooth muscle (SM). The IO muscle also has a soft pulley, mechanically coupled to the IR pulley and located at the insertion of the IO motor nerve. ⁴ Generalizing from the structure of the rectus EOMs and the IO, it might be expected that the SO OL would also insert on connective tissues. However, the SO has been known from antiquity to have a rigid trochlea, into which it would be mechanically impossible for an EOM layer to insert. The relationship of the SO to surrounding connective tissues has not been described. 

The present study was conducted to clarify the function of the OL of the SO, by examining the anatomic relationships of the OL and GL of the SO muscle in monkeys and humans, estimating the total number of SO fibers in monkeys and humans, and estimating the number of fibers in each layer in monkeys. 

Methods

Subjects and Magnetic Resonance Imaging

Ten adult human volunteers were recruited by advertisement and gave written, informed consent, according to a protocol conforming to the Declaration of Helsinki and approved by the Human Subject Protection Committee at the University of California, Los Angeles. High-resolution, T₁-weighted magnetic resonance image (MRI) was performed with a 1.5-T scanner (Signa; General Electric, Milwaukee, WI) scanner. Crucial aspects of this technique, described in detail elsewhere, include use of a dual-phase surface coil array (Medical Advances, Milwaukee, WI) to improve the signal-to-noise ratio and fixation targets to avoid motion artifacts. ^{4 9 10} Some imaging was enhanced with the diffusible paramagnetic contrast agent gadodiamide, given intravenously at 0.1 mmol/kg. ^{10 11} Initially, a localizer axial scan was obtained at 3-mm thickness with a 256 × 192 matrix over a 10-cm² field of view. Contiguous MRIs along the axis of the right orbit were obtained with a 2-mm slice thickness, with a 256 × 256 matrix over an 8-cm² field of view, giving a pixel resolution of 313 μm. Digital MRIs were transferred to Macintosh computers (Apple Computer, Cupertino, CA), converted into 8-bit tagged image file format (TIFF) using locally developed software, and quantified using the program NIH Image (W. Rasband, National Institutes of Health; available by ftp from zippy.nimh.nih.gov or on floppy disc from NTIS, 5285 Port Royal Road, Springfield, VA 22161, part no. PB95500195GEI). 

En Bloc Tissue Preparation

Whole human orbits were obtained from cadavers, in conformity with legal requirements and the guidelines of the Declaration of Helsinki for research involving human tissue. Two human orbits, aged 17 months and 57 years, were obtained from a tissue bank (IIAM, Scranton, PA), from heads fresh frozen shortly after death. The frozen heads were slowly thawed in 10% neutral buffered formalin and fixed in situ within the cranial bones. At autopsy conducted in conformity with legal requirements, a fresh adult human SO was removed, in continuity with its trochlea and associated connective tissues. This specimen was elongated and attached to a cardboard backing during fixation in 10% neutral buffered formalin before embedding en bloc for longitudinal sectioning. 

Two 7-year-old macaque monkeys, one rhesus (Macaca mulatta) and one fascicularis (M. fascicularis), and a capuchin (Cebus apella) monkey were studied. All monkeys were handled in a protocol approved by institutional animal welfare committees and in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Monkeys were euthanatized by intravenous barbiturate overdose and immediately perfused with warm saline followed by neutral buffered paraformaldehyde (4%). The heads were postfixed for several days in 10% neutral buffered formalin. 

Histologic Processing

The fixed orbits were removed en bloc in continuity with the eyelids and orbital bones. With care taken to keep the periorbita intact, the orbital rims and walls were carefully thinned under magnification by using a high-speed drill and rongeurs before decalcification for 24 hours at room temperature in 0.003 M EDTA and 1.35 N HCl. ^{3 4}  

Fixed tissues were dehydrated in graded solutions of alcohol and chloroform or xylenes, embedded in paraffin, and serially sectioned in the coronal plane at 10-μm thickness with disposable metal blades on a microtome (Zeiss HM325; Carl Zeiss Meditec, Thornwood, NY), as previously described. ^{3 4 12} This provided 2800 to 4800 sections per orbit, depending on size. Alternate sets of five contiguous sections were saved and mounted on 50 × 75-mm albumin fixative-coated glass slides. Masson’s trichrome stain was used to indicate muscle and collagen and van Gieson’s stain to indicate elastin. ¹³  

Analysis

In three high-quality sections of each orbit, exact counts were made with a handheld digital counter of all fibers of the OL and GL of the SO muscle at mid orbit. Consistent with earlier work suggesting uniformity of fiber counts at mid orbit of the rectus EOMs, ² sections were selected at 2-mm intervals in monkeys and at 3-mm intervals in humans. No sampling or approximations were used. To obtain an estimate of accuracy, all counts were performed in duplicate for each EOM in all sections. Data reported are the initial counts obtained. Repeatability was ascertained by comparison of the second count with the first. Over all specimens, repeatability of counts in the GL ranged from 0.1% to 0.8%, and in the OL from 0.07% to 1.3%. Repeatability of the total EOM fiber counts ranged from 0.01% to 0.7%. 

Estimates of fiber diameters were obtained from color images obtained with a digital camera mounted on a microscope (BH-2; Olympus, Lake Success, NY). Camera resolution was 3008 × 1960 pixels (DX-1; Nikon, Tokyo, Japan) in 24-bit color. Micrographs were spatially calibrated by imaging a reticule etched on a glass slide. Fibers were manually outlined for analysis with the program ImageJ (developed by Wayne Rasband, NIH). ¹⁴ The area A of each EOM cross-section was determined, and each cross-section was fitted to an ellipse, with its major and minor axes taken as measures of greatest and least diameters of each fiber. Overall mean fiber diameter (d) was computed from cross-sectional area (A) with the assumption of a circular cross section: d = 2(A/π)⁻². 

Results

Magnetic Resonance Imaging

The SO muscle path was demonstrated by MRI in 10 living humans, in whom the findings were uniform and typified by the images in Figure 1 . The SO belly in the deep orbit had a dark signal on T₁ imaging, contrasting strongly with the bright orbital fat surrounding it. In each subject, it was possible in axial images to resolve separately the central tendon and the sheath near the trochlea (Fig. 1A) . The tendon was located centrally and had a bright signal on T₁ imaging, whereas the sheath was peripheral and had a darker signal. Coronal imaging of this same region demonstrated the trochlea and reflected tendon coursing from it inferior to the superior rectus (SR) muscle toward the scleral insertion (Fig. 1B) . The trochlea had a dark signal on T₁ imaging. 

Histologic examination of serially sectioned human orbits permitted correlation with MRI findings in living humans (Fig. 2) . These sections were examined throughout the length of the SO, although space permits publication of only a small sample of micrographs. With Masson’s trichrome stain, SO fibers in the belly stained red, collagenous connective tissues stained blue, and nerves stained pink. At many but not all points along the length of the human SO belly, it was possible to distinguish the OL from the GL (Figs. 2A 2C 2D) . Fibers in the GL were larger and stained bright red, whereas fibers in the OL were smaller and stained darker. In the deep orbit 26.8 mm posterior to the corneal surface, the human SO had minimal investiture by connective tissue and exhibited a C-shaped OL, predominantly on its orbital surface. The smaller and darker-staining OL fibers, in higher magnification on the upper right in Figure 2D , were readily distinguishable from the larger and brighter red GL fibers on the lower left. No transition zone between OL and GL was evident in the deep orbit. In the region of the arborization of the trochlear nerve (Fig. 2B) , the GL and OL were less clearly distinguishable on the basis of fiber size and color, although smaller, darker fibers remained predominant on the periphery (Fig. 2E) . The trochlear nerve (pink, denoted by arrowheads in Fig. 2B ) entered 23.3 mm posterior to the corneal surface, and arborized extensively within the SO belly, but the belly had minimal investiture by connective tissue. Near the point of maximum SO cross section 15.9 mm posterior to the corneal surface (Fig. 2C) , the GL and OL were again readily distinguishable, although precise demarcation of their border was made difficult by transitional fiber characteristics in that area. Near this level, a circumferential investiture of the SO belly by blue-staining collagen developed, into which the peripheral OL fibers became embedded (Fig. 2C) . Actual insertion of OL fibers into the collagenous sheath was evident in higher-power views, such as Figure 2F , illustrating bundles of individual muscle fibers entering the dense collagen of the sheath. At even higher power in Figure 2I , van Gieson’s stain demonstrates black-staining elastin fibrils joining individual OL fibers to the surrounding collagen of the SO sheath, as well as elastin fibrils embedded within the SO sheath. Examination of sections demonstrated that OL fibers did not terminate on the SO tendon and that GL fibers did not terminate on the SO sheath that became increasingly thick as the EOM coursed anteriorly. The central GL fibers coursed farther anteriorly to become contiguous with the bright blue-staining SO tendon 7.7 mm posterior to the corneal surface (Fig. 2G) , encircled by the SO sheath, which stained a distinctive blue-gray color. Both tendon and sheath then passed through the trochlea, a shark-tooth–shaped structure of cartilage anchored to the bony orbit (Fig. 2H) . After passing through the trochlea, the tendon and sheath were reflected laterally and posteriorly where the compact SO tendon broadened and separated from the sheath to course inferior to the SR pulley to insert superotemporally on the sclera (Fig. 2J) . Figure 2Jis a section roughly longitudinal to the path of the reflected tendon. The SO sheath became contiguous with the dense, identically blue-staining collagen of the SR pulley (Fig. 2J) , while the tendon at its scleral insertion was colored identically with the sclera. 

The laminar anatomy of the human SO was confirmed by serial sectioning in the longitudinal plane of a specimen in continuity with the trochlea, removed fresh at autopsy. Serial sections were stained with Masson’s trichrome. A low-power view in Figure 3 (bottom) shows focal attachment of OL fibers to the SO sheath that more anteriorly became coaxial around the central tendon. The central GL fibers continued anteriorly past this OL insertion, where they became contiguous with the central tendon. A higher-power view at the top of Figure 3confirmed that the peripheral SO fibers end in peripheral collagen. 

Histologic findings in the monkey specimens were similar to those in the human, but superior tissue preservation afforded by perfusion revealed additional details. As may be seen in Figure 4A 4B 4C 4D , the demarcation between the OL and GL in the monkey SO was quite clear throughout the length of the EOM. There was minimal investiture of the SO belly by collagen deep in the orbit 23.2 mm posterior to the corneal surface near the entry of the trochlear nerve (Fig. 4A , arrowheads). The GL was central in the EOM, and the OL was peripheral, although not always uniformly concentric. At higher power, the abrupt transition between the larger and brighter-staining GL fibers (Fig. 4C , left) could readily be discerned from the smaller, darker OL fibers (Fig. 4C , right). More anteriorly, 17.4 mm posterior to the corneal surface, the SO belly became surrounded by a collagenous sheath into which inserted the smaller, more darkly staining OL fibers (Fig. 4B) . A high-magnification view illustrates the peripheral OL fibers individually entering deeply into the dense collagen of the sheath (Fig. 4F) . This OL insertion was evident in both monkey specimens. Fibers of the GL became contiguous with the central SO tendon. This transition to central tendon is seen to be distinct from the thick encircling sheath 12.8 mm posterior to the corneal surface, with an intervening gap visible both at low (Fig. 4E)and high (Fig. 4F)power. In the monkey, both the cartilaginous trochlea and central GL fibers extended farther laterally than in the human, but both were reflected in the trochlea (Fig. 4G) . Only GL, and not OL, fibers were evident lateral to the trochlea. As in human specimens, in monkey the SO sheath also inserted on the nasal aspect of the SR pulley whose nasal border is marked by the arrowheads. The SR pulley itself was in close relationship with the trochlea (Fig. 4G) . As in humans, the monkey SO tendon flattened as it coursed inferior to the SR pulley to insert superolaterally on the sclera. Longitudinal sectioning of a monkey SO and associated tissues confirmed passage of the sheath over the trochlea and entry of OL fibers into the collagen of the sheath (Fig. 5) . 

Fiber Numbers

For quantification of fibers, the region of maximum cross-section of the SO belly was determined from inspection of serial sections. Counting was performed in the section exhibiting maximum cross-sectional area and repeated in one section 3 mm anterior and another 3 mm posterior to the section containing the maximum area in human, and in one section 2 mm anterior and another 2 mm posterior in monkey. These values were then averaged for each orbit. Absence of a clear demarcation between the OL and GL in the human SO made it impossible to obtain a reliable count of the two layers separately, and so only total counts were obtained for the entire SO in humans. The mean (±SD) number of fibers in the 17-month-old human SO was 14,353 ± 614 (n = 3 sections), whereas in the 57-year-old specimen it was 19,180 ± 1,946. 

The two layers were individually counted in the monkey specimens (Table 1) , although the total number of fibers was no more than half that of humans. The fascicularis monkey had 6949 ± 231 total fibers, whereas the rhesus monkey had 7393 ± 1089. In each monkey, approximately 61% of fibers were in the GL and 39% in the OL. 

Fiber Size

Whereas the border between the OL and GL of the human was not sufficiently precise for fiber counting, regions of the two layers remote from the border could unequivocally be defined for analysis of fiber size. The striking qualitative impression of smaller fibers in the SO OL was validated quantitatively. Coronal sections of excellent quality were selected from the mid orbital regions of the SO, with care taken to avoid obliquity of the plane of section to the fiber path. From histologic sections, borders were digitally outlined by hand for the cross sections of four sets each of 20 contiguous fibers in the OL and four sets each of 20 contiguous fibers in the GL, with the OL and GL samples in each case paired to represent similar EOM regions around the SO circumference. Measurements included large, presumably singly innervated fibers, as well as small, presumably multiply-innervated fibers. The area of each EOM cross section was determined, and each cross section was fitted to an ellipse with major and minor axes that were taken as measures of greatest and least diameters of each fiber. In the 17-month-old human, the mean ± SD GL cross-sectional area was 161 ± 42 μm², significantly greater than the mean OL cross-sectional area of 85.8 ± 19.3 μm². The mean greatest diameter of GL fibers was 16.1 ± 2.4 μm, whereas the mean greatest diameter of OL fibers was 12.3 ± 2.7 μm. The mean least diameter of GL fibers was 12.5 ± 2.1 μm, whereas the mean least diameter of OL fibers was 8.8 ± 1.3 μm. Based on the assumption of a circular cross section, the mean overall diameter for GL fibers was 14.2 ± 1.9 μm, whereas the mean overall diameter for OL fibers was 10.4 ± 1.2 μm. For the pooled data, all these differences were significant at P < 10⁻²¹, but all differences were also significant at P < 0.00001 in each of the four subsamples. The coefficient of variation (CV) of overall mean diameter for GL fibers was 0.13, whereas the CV for OL fibers was 0.12. 

Fiber diameters were similarly determined in the rhesus monkey specimen, but from subsamples containing 20 fibers from the OL and GL in each of seven representative regions of the SO at mid orbit. Fibers in the rhesus monkey GL varied considerably in size and in staining characteristics. In the monkey, mean (± SD) GL cross-sectional area was larger than in the human at 384 ± 209 μm², significantly greater than the mean OL cross-sectional area of 132 ± 51 μm². The mean greatest diameter of GL fibers was 26.0 ± 8.1 μm, whereas the mean greatest diameter of OL fibers was 15.6 ± 3.2 μm. The mean least diameter of GL fibers was 17.6 ± 5.5 μm, whereas the mean least diameter of OL fibers was 10.5 ± 2.4 μm. Based on the assumption of a circular cross section, the mean overall diameter for GL fibers was 21.2 ± 6.2 μm, whereas the mean average diameter of OL fibers was 12.7 ± 2.2 μm. For the pooled data, all these differences were significant at P < 10⁻²⁷, but all differences were also significant at P < 0.001 in each of the seven subsamples. The CV of overall mean diameter of monkey GL fibers was more than double the size in human at 0.29, whereas the CV of monkey OL fibers was moderately higher than in human at 0.17. 

Discussion

The gross anatomy of the human SO has long been understood from classic cadaver studies. The SO originates near the orbital apex just above the annulus of Zinn, and gradually becomes tendinous posterior to the trochlea. Reflecting in its pulley the trochlea, the SO tendon courses posteriorly and temporally inferior to the SR muscle to insert on the sclera lateral to the SR muscle. ¹⁵ In the present study, this gross anatomy was readily demonstrable by MRI in living humans, but the relationship of the SO tendon to its sheath was also evident on MRI. Histologic examination demonstrates that in the deep orbit, the SO muscle is only minimally invested by connective tissues that become progressively more developed anteriorly as the EOM fibers transition to dense tendon. A fibrovascular sheath has previously been described to surround the intratrochlear SO tendon, with a bursalike structure lying between the sheath and the cartilaginous trochlear saddle. ¹⁵ On T₁-weighted MRIs performed in the present study, the more vascular central SO tendon appeared brighter, whereas its less vascular sheath appeared darker. The sheath was readily demonstrable on MRI, permitting its noninvasive study in normal and pathologic states. 

In the rectus EOMs, the C-shaped OL on the orbital surface was readily distinguished from the central GL on the basis of smaller fiber size and darker red staining in the former and the larger, bright red fibers in the latter. ^{2 11} The microscopic appearance of rectus EOMs is similar in both humans and monkeys. ^{2 11} In contrast, the SO muscle has been described as more concentrically organized, with the OL encircling most of the EOM’s periphery, and the GL occupying the core. ¹⁶ Although in the present study the OL of the SO muscle was clearly demarcated from the central GL in the monkey, the absence of such a clear border in the human was consistent with the gradual transition zone described by Spencer and Porter, ¹⁶ containing an admixture of fiber types from either zone. 

The SO contains fewer fibers than the rectus EOMs. In the present study the human SO was found to contain 14,400 to 19,200 fibers, whereas the monkey SO contained 7,000 to 7,400. In comparison, the human rectus EOMs contain 17,700 to 24,500 total fibers, whereas the monkey rectus EOMs contain 8,000 to 11,800 fibers. ² However, the size of the SO fiber was similar to that of rectus EOMs. In the present study, the mean GL fiber diameter of the human SO was 14.2 ± 1.9 μm; the corresponding mean diameter of the OL was significantly smaller at 10.4 ± 1.2 μm. These dimensions are virtually identical with those of human rectus EOMs. It has been reported that the mean GL fiber diameter of the human medial rectus (MR) muscle in this 17-month-old specimen was 15.0 ± 4.0 μm, and the mean OL diameter was 10.8 ± 2.7 μm. ²  

In the monkey the SO GL contained 4300 to 4600 fibers, whereas the rectus GLs contained 4400 to 5900. The SO OL contained 2700 to 2800 fibers, whereas rectus OLs contained 3700 to 5900. Thus, in monkeys, the GL of the SO contains only slightly fewer fibers than that of the rectus EOMs, whereas the OL contains approximately one half to two thirds the corresponding number of fibers in the rectus EOMs. In the SO, approximately 60% of total fibers are in the OL versus the comparable figure of 39% to 53% in monkey rectus EOMs. ² The relative proportion of OL fibers in the monkey SO was similar to that of rectus EOMs. Although less numerous, SO fibers in the monkey were substantially larger and more variable in size than comparable fibers in the human SO. The mean overall GL fiber diameter in the monkey was 21.2 ± 6.2 μm, 49% larger than in the human. Mean overall OL fiber diameter in the monkey of 12.7 ± 2.2 μm was only 12% larger than the comparable value in the human, suggesting that these fiber size differences between monkey and human are biologically significant rather than simply artifacts of differences in postmortem conditions and fixation. 

The myosin content, mitochondrial content, and vascularity of rectus EOM OL fibers is tailored to their continuous activity to maintain the tension required to position the pulleys against their elastic suspensions. ^{6 16} Based on morphologic similarity between OL fibers in rectus EOMs and the SO in both human and monkey, the present study suggests that OL fibers of the SO may also fulfill a role in maintenance of tonic tension to control pulley positions. Possible electrophysiological correlates of this anatomy have yet to be investigated. 

It is now recognized that the OLs of the rectus and IO EOMs contain a substantial proportion of total fibers and in each case serve to regulate EOM pulling directions via insertions on the connective tissues of the pulley system. ^{3 4 8 17} The present study extends this theme to the SO. Both in human and in macaque monkey specimens, fibers of the SO OL were arranged around the periphery of the central GL. Whereas the larger and more pleomorphic GL fibers coursed anteriorly to become contiguous with the central SO tendon, the smaller and more uniform OL fibers inserted on the SO sheath. The SO sheath thickened anteriorly as the OL terminated and reflected in the trochlea along with the SO tendon. Another novel finding of this study is that, distal to its reflection, the SO sheath terminated on the nasal aspect of the SR pulley, whereas the SO tendon inserted, as classically described, on the sclera. 

In coronal sections, OL fibers have been found to be entirely absent in the terminal anterior portions of rectus EOMs in both monkeys and humans, confirming the results of longitudinal sectioning indicating that the OL fibers do not insert on the sclera. ^{2 3} The present study demonstrated that the SO OL inserts on the SO sheath in humans (Figs. 2F 2I)and monkeys (Fig. 4F) . The nature of this insertion is similar to that of the rectus EOMs ^{4 12} and the IO. ^{3 8} In addition, the SO sheath is reflected through the trochlea to insert on the nasal aspect of the SR pulley (Fig. 2J) . This arrangement permits contraction of the SO OL to translate the SR pulley nasally, similar to the way that contraction of the IO OL translates the IR and LR pulleys. ⁴  

Insertion of the SO sheath on the SR pulley provides a mechanism for regulation of EOM pulling direction by the SO OL: contraction of the OL would shift the SR pulley nasally, and relaxation of the OL would permit the SR pulley to shift temporally. The OL of the IO is known to insert on both the IR and LR pulleys in a manner that would produce extorsional shift in their positions during IO contraction, although this insertion degenerates over the human lifespan. ^{4 17} Torsional reconfiguration of the rectus pulley array has multiple kinematic implications, ⁶ most of which can be understood in the context of Listing’s law (LL) of ocular torsion. Maintenance of appropriate anteroposterior positioning of rectus EOM pulleys allows the rectus EOM pulling directions to change by half of the eye position during pursuit, saccades, and fixations with the head motionless. ⁶ The half-angle dependence of rectus pulling direction on eye position is consistent with LL, ¹⁸ a quantitative description of ocular torsion that includes a corollary and original statement that any eye orientation can be reached from a primary position by rotation about a single axis lying in Listing’s plane. ¹⁹ The half-angle dependence of the rectus EOM pulling direction also makes the peripheral ocular motor apparatus appear commutative to the brain in the mathematical sense in that the ultimate eye position has no significant dependence on the sequence of rotations required to reach it. ^{18 20 21 22} If rectus pulleys were fixed in torsional positions relative to the orbit, globe torsion would impart a new torsional action to each of the rectus EOMs, ²³ complicating ocular motor control. Because the rectus pulleys are located as far posterior to globe’s center as the insertions are anterior to the globe’s center and are subject to the trigonometrically small angle approximation, pulleys that did not move at all during ocular torsion would cause the rectus EOM’s pulling directions to tilt anteroposteriorly by half of the ocular torsional angle. This half-angle dependency would be analogous to LL behavior for ocular torsion. 

Listing’s law does not prevail during head tilt. During static head tilt, MRI demonstrates that the rectus pulley array shifts torsionally in the same direction as ocular torsion (Demer JL. IOVS 2003;44:ARVO E-Abstract 2736), correlating with a parallel offset in the torsional position of Listing’s plane in which the rotational axes of saccades and pursuit remain. ^{24 25 26 27 28} However, MRI during static head tilt suggests there is a torsional shift of the rectus pulley array equal to half of ocular torsion. This torsional pulley shift would cause each rectus EOM’s pulling direction to change by one fourth the change in globe torsion. ⁶ During head rotations, the vestibulo-ocular reflex (VOR) velocity axis rotates by 0% ^{29 30} to 25% ³¹ of the ocular angle, contrasting with the 50% rotation required by LL. Recent experiments in humans undergoing transient whole-body yaw at high acceleration in darkness has supported a quarter angle dependence of the VOR velocity axis on vertical eye position. ³² The torsional shift in the rectus pulley array half of ocular torsion is consistent with the known quarter angle dependence of the VOR on torsional eye position. The action of the OL of the SO on the coronal plane position of the SR pulley may therefore be an important factor in the regulation of VOR kinematics. 

Another situation violating LL is convergence, in which the globe’s extorsion is consistently observed for central gaze in both humans ^{33 34 35 36 37} and monkeys. ³⁸ During asymmetrical accommodative- or disparity-induced vergence, this temporal rotation occurs in both the aligned and converging eyes, independent of eye position. ³⁹ Thus, the classic form of LL is not observed for near viewing. The Listing plane for each eye tilts temporally with convergence, ^{33 34 39 40 41} in a manner that has been described as the binocular extension of LL. ⁴² In humans, MRI during asymmetrical convergence is associated with an extorsion of the rectus pulley array and with evidence of SO relaxation. ¹⁰ Single-unit recordings in monkey trochlear nucleus show a greater reduction in activity in convergence than in conjugate adduction, prompting Mays et al. ⁴³ to propose that SO relaxation mediates excyclotorsion in convergence. The present finding that the OL of the SO inserts on the SR pulley explains at part of the extorsion of the rectus pulley array in convergence. The known insertion of the OL of the IO on the inferior (IR) and lateral rectus (LR) pulleys, ¹⁷ as well as MRI evidence of IO contraction in convergence, ¹⁰ may explain the repositioning of the IR and LR pulleys. 

In conclusion, the primate SO is similar to the other five oculorotary EOMs, in that it has similar sized fibers and a distinct OL containing approximately 40% of total EOM fibers. Like the rectus and IO EOMs, the SO OL inserts on the connective tissue pulley system and so appears to control the direction of application of EOM force. Although the direction of application of SO GL force is determined by the rigid trochlea, the SO OL influences the direction of application of rectus EOM forces in a manner that may have broad kinematic implications for the VOR and for convergence. This insight extends the concept of active control of pulley positions to include a contribution from the SO muscle. 

 Supported by National Eye Institute Grant EY-08313 (JLD). JLD is the recipient of an unrestricted award from Research to Prevent Blindness and is the Grace and Walter Lantz Professor of Ophthalmology.

 Submitted for publication September 27, 2004; revised April 8, 2005; accepted April 21, 2005.

 Disclosure: R. Kono, None; V. Poukens, None; J.L. Demer, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(A) An axial MRI of 2-mm thickness obtained with gadodiamide contrast showing SO tendon (SOT) reflecting in the trochlea. Note separate resolution of the darker sheath surrounding the brighter SO tendon. (B) Noncontrast coronal MRIs of 2-mm thickness, arranged from posterior to anterior skipping 2 mm between the planes displayed.

Figure 2.

 

Series of coronal histologic sections of 17-month-old human SO muscle, with Masson’s trichrome stain, except for van Gieson’s elastin stain in (H). (A) Posterior to the corneal surface by 26.8 mm. The SO muscle belly near its origin in the deep orbit shows a C-shaped OL, mainly on the orbital surface of the more central GL. (B) Posterior to the corneal surface by 23.3 mm. The trochlear nerve (pink, arrowheads) entered and arborized extensively within the SO belly. The OL and GL were difficult to distinguish in this region. (C) Posterior to the corneal surface by 15.9 mm. Approximately the maximum cross sectional area of the SO. The more central GL fibers were brighter red than the OL fibers, but the precise demarcation between the layers was unclear. (D) Higher-power view of boxed area in (A). The GL fibers (left) were brighter red and larger than The OL fibers (right). (E) Higher-power view of boxed area in (B). The GL fibers (right) were brighter red and larger than the OL fibers (left). (F) Higher-power view of boxed area in (C) showing insertion of the OL fibers into the dense collagen (blue) of the peripheral sheath. (G) Posterior to the corneal surface and posterior to the trochlea. Central bright blue fibers of the SO tendon were surrounded by the gray-blue collagen of the peripheral sheath. (H) Posterior to the corneal surface by 4.8 mm, through the shark-tooth–shaped cartilage of the trochlea. Note that the blue-staining SO tendon was sectioned both within the trochlea (right, large arrowhead) and in its reflected portion (left, small arrowhead). The remnant of the orbital bone at lower right stained bright red. (I) Higher-power view of boxed area in (F). Van Gieson’s stain demonstrated black elastin fibrils bridging the dark-staining OL fibers and the red-orange staining collagen of the sheath, within which were also embedded elastin fibrils. (J) Posterior to the corneal surface by 10.1 mm. At this anterior level, the thin, compact SO tendon, stained dark blue, coursed inferior to the SR pulley to insert on the sclera. The SO sheath inserted on the nasal aspect of the dense blue-staining collagen of the SR pulley.

Figure 3.

 

Human SO muscle sectioned longitudinally and stained with Masson’s trichrome. Bottom: OL fibers present at the periphery in the posterior orbit (left) terminated on a focal connective tissue aggregation within the rectangle that in turn was attached to the peripheral SO sheath that coursed anteriorly (left) to surround the tendon at the core of the complex. GL fibers in the center continued farther anteriorly past the OL insertion to become contiguous with the core tendon. Bottom: The higher-power view of the region in the rectangle illustrates individual OL fibers inserting into the connective tissue of the sheath.

Figure 4.

 

Coronal sections of 7-year-old monkey orbit in the region of the SO muscle, stained with Masson’s trichrome. (A) Posterior to the corneal surface by 23.2 mm. The trochlear nerve entered the SO superonasally, with intramuscular arborizations staining pink (arrowheads). Note the minimal collagenous sheath encircling the SO. (B) Posterior to the corneal surface by 17.4 mm. The SO sheath was increasingly developed anteriorly. The OL fibers inserted into the dense, blue-staining collagen of the SO sheath. (C) Higher-power view of boxed area in (A). Larger, bright red GL fibers at left were easily distinguished from the smaller and darker OL fibers at right. (D) Higher-power view of the boxed area in (B). The OL fibers at left inserted into the blue-staining collagen of the SO sheath. (E) Posterior to the corneal surface by 12.8 mm. The transition from the SO muscle to the tendon was surrounded by a well-developed sheath. (F) Higher-power view of boxed area in (E). Insertion of OL fibers into the fibrovascular sheath (left) is demonstrated. (G) Posterior to the corneal surface by 4.6 mm. The more proximal distal SO tendon, encircled by its sheath (left), was reflected over the extensive, cartilaginous trochlea (Tr) that appeared twice in this section, due to its curved shape. The distal SO is demonstrated at left. The SR muscle was encircled by its dense collagenous pulley. The triangular extension of the SR pulley on the nasal side (arrowheads) was traced in serial sections (not shown), to be in continuity with the reflected SR sheath.

Figure 5.

 

Longitudinal micrographs of monkey SO muscle and trochlea, stained with Masson’s trichrome. (A) Low-power view shows the SO belly (left) and the trochlea (right). Note the blue-staining collagen of the sheath reflecting over the crescent-shaped cartilage of the trochlea. (B) Higher-power view of the surface region near the asterisk in (A), showing the blue collagen sheath on the SO surface. (C) Very high-power view shows superficial SO fibers entering into the collagen of the sheath.

Table 1.

 

Number of Fibers in Monkey Superior Oblique Muscle

Species	Global Layer	Orbital Layer	Total
Fascicularis	4,434	2,254	6,688
	4,229	2,900	7,129
	4,117	2,912	7,029
Mean ± SD	4,260 ± 161	2,689 ± 376	6,949 ± 231
Rhesus	2,951	3,217	6,168
	5,547	2,214	7,761
	5,166	3,084	8,250
Mean ± SD	4,555 ± 1,402	2,838 ± 545	7,393 ± 1,089

 Counts of fibers in the orbital and global layers of the superior oblique muscle of two monkeys obtained in three different sections at mid orbit in each specimen, spaced at 2-mm intervals.

The authors thank Nicolasa de Salles and Frank Henriquez for technical assistance and Joel M. Miller, Lawrence Tychsen, and James Lynch for generously providing monkey specimens.