Article Text

Download PDFPDF

Decussating axons segregate within the anterior core of the primate optic chiasm
  1. Jonathan C Horton,
  2. Mikayla D Dilbeck,
  3. John R Economides
  1. Department of Ophthalmology, University of California San Francisco, San Francisco, California, USA
  1. Correspondence to Jonathan C Horton, Department of Ophthalmology, University of California San Francisco, San Francisco, California 94143, USA; Jonathan.Horton{at}


Background The axons of ganglion cells in the nasal retina decussate at the optic chiasm. It is unclear why tumours cause more injury to crossing nasal fibres, thereby giving rise to temporal visual field loss in each eye. To address this issue, the course of fibres through the optic chiasm was examined following injection of a different fluorescent tracer into each eye of a monkey.

Methods Under general anaesthesia, cholera toxin subunit B—Alexa Fluor 488 was injected into the right eye and cholera toxin subunit B—Alexa Fluor 594 was injected into the left eye of a single normal adult male rhesus monkey. After a week’s survival for anterograde transport, serial coronal sections through the primary optic pathway were examined.

Results A zone within the core of the anterior and mid portions of the optic chiasm was comprised entirely of crossing fibres. This zone of decussation was delineated by segregated, interwoven sheets of green (right eye) and red (left eye) fibres. It expanded steadily to fill more of the optic chiasm as fibres coursed posteriorly towards the optic tracts. Eventually, crossed fibres became completely intermingled with uncrossed fibres, so that ocular separation was lost.

Conclusions A distinct, central compartment located within the anterior two-thirds of the optic chiasm contains only crossing fibres. Sellar tumours focus their compressive force on this portion of the structure, explaining why they so often produce visual field loss in the temporal fields.

  • Optic Nerve
  • Visual pathway

Data availability statement

Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information. See above responses.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


In Opticks, Newton proposed that a hemidecussation of fibres occurs at the optic chiasm.1 2 His idea was relegated to an appendix, consisting of 15 queries designed to stimulate ‘further search to be made by others’. Many subsequent studies have labelled a single eye to show the crossing of nasal fibres, but no study has labelled the fibres from both eyes.3–12 The aim of this report is to describe, by means of independent tracers, how retinal ganglion cell axons emanating from each eye dovetail as they traverse the optic chiasm.

Materials and methods

Tracer eye injections

This experiment adhered to a protocol approved by the University of California San Francisco Institutional Animal Care and Use Committee and was conducted in accordance with the Association for Research in Vision and Ophthalmology Statement on the Use of Animals. An adult male macaque (Macaca mulatta) was placed under general anaesthesia with ketamine hydrochloride (15 mg/kg intramuscularly). Both pupils were dilated with 1% cyclopentolate HCl and 2.5% phenylephrine HCl. Through closed eyelids the globes were massaged gently for 10 min to reduce the intraocular pressure. Topical 1% proparacaine HCl was applied to the conjunctival sacs for local anaesthesia. An injection of 500 µg of cholera toxin subunit B conjugated to Alexa Fluor 488 (Thermo Fisher Scientific) dissolved in 50 µl of sterile saline containing 5% dimethyl sulfoxide was made into the right eye.13 Immediate examination with an indirect ophthalmoscope showed a plume of tracer in the mid-vitreous with no injury to the eye. The identical procedure was then performed in the left eye, except that the tracer was cholera toxin subunit B conjugated to Alexa Fluor 594.

Histological processing

After 7 days to allow transport of the anterograde tracers, the animal received a lethal dose of pentobarbital (150 mg/kg). Transcardial perfusion was performed with 1 L of 0.9% saline followed by 1 L of 4% paraformaldehyde dissolved in 0.1 M phosphate buffer. A tissue block comprised of the primary optic pathway from the optic discs to the lateral geniculate nuclei was placed into 4% paraformaldehyde in 0.1 M phosphate buffer with 30% sucrose for cryoprotection. After several days, serial coronal sections were cut at 60 µm on a freezing microtome. They were coverslipped with 80% glycerin/20% Tris buffer, 0.1 M, pH 8.5.

Image analysis

Tissue sections were examined in a Zeiss Axiophot microscope using the Chroma ET-GFP-Cy2-FITC filter set for Alexa Fluor 488 and the Chroma ET-CY3-TRITC filter set for Alexa Fluor 594. Sections were photographed through each filter set and the two images were imported into layers in Photoshop (Adobe). Global adjustments were made to the brightness and contrast of each image. Once the red and green signals were balanced, the Difference Blend mode was applied to combine the layers into a single image.

A drawing of the optic chiasm compressed by a pituitary adenoma was made by a professional medical artist. The case chosen was that of a middle-aged patient with a classic bitemporal hemianopia revealed by automated threshold perimetry. T1-weighted gadolinium-enhanced axial and coronal MRIs (1.2 mm slice thickness) were used to prepare the figure. The optic nerves, chiasm and tracts in serial coronal image were outlined as well as the contours of the pituitary tumour. These outlines were imported into Igor Pro (WaveMetrics) to create a three-dimensional, rotatable volume, which was consulted during the process of preparing the illustration. The UCSF Institutional Review Board exempted use of MRIs from this single subject from human studies approval.


Figure 1 shows a series of coronal sections cut through the optic pathway. Despite its name, the primate optic chiasm forms an ‘H’, not an ‘X’. In the macaque, the crossbar of the ‘H’ is very short, measured by the separation of the optic nerves just anterior to the chiasm. In this specimen, it was only 750 µm, compared with 4.7 mm in the human.14

Figure 1

Series of coronal sections beginning from the proximal optic nerves (−180 μm) through the optic chiasm (0–2820 μm) and then 9 mm further posteriorly at the level of the lateral geniculate nuclei, showing at each stage the axons serving the right eye (Alexa Fluor 488, green) and the left eye (Alexa Fluor 594, red). Where fibres from the two eyes are finely interspersed, the fluorescent signal appears yellow. Note the suprachiasmatic nuclei labeled yellow in the section at 1800 μm.

Each optic nerve contained a pure population of axons. No label was present from the tracer that was injected into the opposite eye (figure 1, −180 µm). Even at the point where the optic nerves became conjoined (figure 1, 0 μm), they remained free of axons from the other eye. Wilbrand reported that crossing fibres detour for several millimetres into the proximal contralateral optic nerve before doubling back to enter the optic tract.15 Figure 1 shows that this aberrant fibre loop, known as ‘Wilbrand’s knee’, is not present when both optic nerves are intact. It forms in patients only after severe monocular vision loss that causes damage to the optic nerve.16 Wilbrand’s knee, therefore, is irrelevant to the issue of how lesions that compress the primary optic pathway give rise to various patterns of visual field loss because it does not exist in normal anatomy.17 18

After the optic nerves united to form the chiasm, axons began to stream across the midline (figure 1, 300 μm). About 30 green fibre leaflets interdigitated with a similar number of red fibre leaflets. The crossing axons gave the impression of being grouped into fascicles that ran laterally within the tissue section. In reality, the axons were arrayed in sheets of fibres, cut obliquely, that swept posteriorly and laterally.

Progressing posteriorly, the crossing axons formed an enlarging stack within the central chiasm (figure 1, 660 μm). The leaflets became more numerous, about 45 for each set, and more finely interwoven. The central zone of crossing fibres was still sharply segregated from the populations of fibres from the left eye and the right eye, some destined to cross but most not, that surrounded it on either side.

By halfway through the optic chiasm, most decussating fibres had journeyed across the midline (figure 1, 1260 μm). As a result, crossing axons were distributed asymmetrically, with the majority concentrated on the side contralateral to their eye of origin. The crossing axons occupied the centre of the chiasm and were fairly well segregated from the flanking contingents of uncrossed axons. However, the flanking zones filled by uncrossed axons were infiltrated by numerous thin bands of crossed axons that extended all the way to the surface of the chiasm.

Further posteriorly (figure 1, 1800 μm) crossed and uncrossed axons became progressively more intermingled, although crossed fibres still predominated in the core of the chiasm. At this level, nearly all nasal fibres had undergone decussation. Consequently, the fibres in each half of the optic chiasm represented the contralateral visual field. The suprachiasmatic nuclei, labelled by axons from both eyes, were visible just above the optic chiasm.19 They appeared yellow because red axon terminals from the left eye and green axon terminals from the right eye were finely interspersed.13

In the posterior optic chiasm (figure 1, 2160 μm) crossed and uncrossed fibres were intermingled extensively, changing the hue of the tissue section from interlaced red and green to an orange or yellow colour. Nonetheless, crossed fibres still predominated slightly in the medial chiasm, reflected by a green tinge to the left paracentral chiasm and a red tinge to the right paracentral chiasm.

In the last section (figure 1, 2820 μm), through the optic chiasm before it divided into the optic tracts, the fibres from the two eyes were nearly completely integrated. They remained this way until they reached the lateral geniculate nuclei, where they segregated again, to generate alternating monocular laminae. This transition was signified by the re-emergence of red and green label in the geniculate image. As expected, the ipsilateral eye occupied laminae 2, 3 and 5 and the contralateral eye filled laminae 1, 4 and 6.

Figure 2 shows the central zone occupied by fibres undergoing decussation within the optic chiasm. At 300 µm behind the junction of the optic nerves, fibres crossing the midline occupied an exclusive core zone that filled 20% of the chiasm. Moving posteriorly the size of this central zone of decussating fibres expanded steadily. It occupied 55% of the cross-sectional area of the mid-chiasm (figure 1, 1260 μm). In the posterior third of the chiasm, crossed and uncrossed fibres became so intermingled that one could no longer reliably assign them to separate zones.

Figure 2

Sketches at various levels of the optic chiasm in this figure, delineating an exclusive, central zone (yellow shading) that contains axons which have undergone decussation. These fibres occupy an expanding percentage of the anterior and middle optic chiasm. This is precisely the zone subject to the greatest compression and distortion from a sellar tumour.


When a single tracer is deposited into one eye to visualise fibres coursing through the optic chiasm, the fibres from the other eye appear as blank, unlabeled regions sandwiched between labelled axon fascicles. We did not expect independent labelling of the fibres from the other eye to add much extra information. Surprisingly, however, the vivid contrast of the fluorescent tracers, the symmetry of their pattern, and the colour shift to yellow where axons interdigitate on a microscopic scale provide additional data that help one to understand how the primate optic chiasm is organised.

The course followed by axons from each eye through the optic chiasm is relevant to a longstanding puzzle: why does compression from a lesion such as a pituitary adenoma cause a bitemporal hemianopia? It should be stressed that sometimes it does not.20 The resulting visual field defect may be incomplete, asymmetrical or reflect mixed compression of the optic nerve(s) and chiasm.16 21 These caveats aside, the stereotypic pattern of visual field loss from chiasmal compression is a temporal defect that worsens abruptly at the vertical meridian. This fact means that axons which decussate are damaged more severely by compressive lesions than those that remain ipsilateral. The reason for this differential vulnerability remains controversial.22–25

Some textbooks show fibres from the nasal retina crossing immediately on entry into the optic chiasm.26 27 This depiction is obsolete, based on the old notion that they sweep across the anterior chiasm to enter the other optic nerve, where they form Wilbrand’s knee. In reality, only a minority of nasal fibres cross in the anterior chiasm, and these are located in the central core of the structure (figure 2, 300 μm). It remains unknown where the remaining nasal fibres, destined to cross but not yet crossed, reside in the anterior chiasm. It is natural to suppose that they lie close to the midline, in anticipation of decussation. However, data from optic tract injections of a retrograde tracer indicate that they can be scattered quite widely, with many located laterally in the chiasm.28 The definitive experiment to elucidate the fibre organisation of the optic chiasm has not yet been carried out. It would combine in the same animal the injection of a different anterograde tracer into each eye along with the injection of a different retrograde tracer into each optic tract. This proposed quadruple-tracer experiment would provide independent labels for the four axon populations: right eye decussating, right eye non-decussating, left eye decussating and left eye non-decussating. Identification of each population tagged by a unique combination of two different tracers would enable one to determine precisely how they are arranged at every level within the optic chiasm. It would also settle another unresolved point: to what extent are fibres originating from the nasal retina aggregated in the medial half of each optic nerve just anterior to the optic chiasm?

The greater susceptibility of decussating axons to damage from compression is explained by their physical location. Within the anterior two-thirds of the optic chiasm, they are concentrated within the central core of the structure (figure 2). The dome of a pituitary tumour arising from the sella turcica will come first into contact with the central chiasm. As the tumour expands, maximum force will be exerted on crossing fibres. Kosmorsky and colleagues have provided direct evidence in support of this supposition by making measurements in cadavers with simulated pituitary tumours.24 Higher pressures were recorded in the central portion than in the lateral edges of the optic chiasm. Crossing central fibres may also be more vulnerable to injury because they are oriented in a perpendicular rather than a parallel fashion, thereby focusing compressive forces at contact points.25

Tumour compression pushes the chiasm upwards, flattening and widening it (figure 3). Elevation is usually maximal in the central chiasm, concentrating pressure in this region. In addition, elevation increases the distance that fibres must travel. The increase is greater for crossed than uncrossed fibres, because the latter skirt around the lateral edges of the chiasm, which undergo less elevation. Pathological stretching of axons is known to cause conduction block, axolemmal disruption and eventual degeneration.29–31 A greater increase in path length for crossed fibres, relative to uncrossed fibres, is likely to be another factor that contributes to bitemporal visual field loss from sellar tumours.

Figure 3

A pituitary adenoma causes elevation and splaying of the optic chiasm. The optic nerves anchor the anterior chiasm, preventing it from moving out of the way of a growing pituitary tumour. The posterior chiasm starts from a higher position, and is able to move further upwards because the optic tracts are more mobile than the optic nerves. As a result, the anterior and middle optic chiasm are subject to greater compression and distortion from a tumour.

In the posterior optic chiasm, the decussation of fibres is largely complete. Axons overlap that serve the nasal and temporal retina, giving the image a yellow colour (figure 2, 2160 μm and 2820 µm). Tumour compression of this portion of the optic chiasm should affect nasal and temporal visual fields equally. Why then, does bitemporal hemianopia predominate? We propose that sellar tumours compress the anterior and middle portions of the optic chiasm more severely than the posterior optic chiasm. In the human, the optic pathway rises at 45° from the skull base, giving the optic chiasm a tilted orientation. The anterior border of the optic chiasm is located 4 mm below its posterior border.32 This means that a growing sellar tumour exerts more pressure on the anterior half than the posterior half of the chiasm.

The optic nerves tether the anterior chiasm, thereby preventing it from escaping a growing sellar tumour. In contrast, the posterior chiasm is joined to the optic tracts, which are relatively free to move upwards to accommodate an expanding tumour. The fact that the optic tracts are located far from the sella turcica, and are able to move out of harm’s way, explains why homonymous hemianopia is rare in patients with pituitary adenoma. In contrast, the optic nerves are situated much lower and are anchored at the optic canals and the anterior chiasm, leaving them more vulnerable. Large pituitary tumours compress the optic nerves, in addition to the optic chiasm. They impinge directly on the medial, inferior aspects of the optic nerves, where crossing nasal fibres are more likely to be prevalent. In many patients, bitemporal field loss reflects a combination of the effects produced by pressure on one or both optic nerves, plus the optic chiasm.

The observations in this report are based on the macaque, not the human. The course of optic nerve fibres through the human optic chiasm has been examined with diffusion tensor imaging.33–38 Unfortunately, the resolution of this technique is too low to show reliably the decussation of nasal fibres, let alone the fine details of intrachiasmal fibre organisation. Despite claims, diffusion tensor imaging is not an adequate substitute for histopathological studies.39 Anatomical studies in monkeys are still required to probe features of primate anatomy that are beyond the resolution of non-invasive imaging.

Data availability statement

Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information. See above responses.

Ethics statements

Patient consent for publication


Kenneth Probst prepared the drawing of the optic chiasm compressed by a pituitary adenoma. Rakesh Nanjappa assisted with fluorescence microscopy.



  • Collaborators Not applicable.

  • Funding This work was supported by grants EY029703 (JCH) and EY02162 (Vision Core Grant) from the National Eye Institute and by an unrestricted grant from Research to Prevent Blindness. The California National Primate Research Center is supported by a Base Grant from the NIH Office of the Director, OD011107.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; internally peer reviewed.

  • Author note This article is published to mark the Festschrift held at the Royal Society of Medicine in London, UK, on 25th and 26th March 2021 to celebrate the retirement of Dr Gordon T Plant . A complete recording of the event is available on the website of the United Kingdom Neuro-Ophthalmology Society: